Welcome back to The Customer Continuum. Issue #37.

When I was early in my career at Marketo, I believed that doing a good job just meant earning more of it. And for a while that was fine, because I wanted to prove I was worth it. So when they asked me to help bring five hundred CMOs to our summit in under ninety days, I said yes. When they layered an executive engagement program on top, and a handful of other things after that, I said yes to those too. I was new, and “sure, let’s do it” felt like the thing that got you taken seriously.

Then the reality of it landed. My team was only so big, and we were already stretched thin, staring down a deadline none of us felt good about. We pulled it off, and the event was amazing, the program too. But it cost us. People started to leave, and morale slid. Then we turned around and spent the next nine months getting ready to do the whole thing over again the following year.

What stayed with me is that we rebuilt almost everything from scratch every time, because we’d never stopped to build the systems underneath. If we could have kept a version of what we made and started the next round at seventy percent instead of zero, that job would have felt completely different. The volume of work was real, but doing all of it from a blank page, year after year, is what wore us down.

That’s the gap I think we can finally close, because the work that used to get rebuilt by hand can now be built once and run by a system. And it changes what a customer marketing team even needs to look like.

For as long as I’ve been in this industry, a complete customer marketing organization falls under seven pillars: community, adoption, voice of customer, advocacy, communications, education, and executive engagement, and the org chart for them has always been a headcount plan. Want more advocacy, you hire an advocacy person. Want a real community, you hire a community manager. The work grew with the number of people, which is why “do more with less” felt impossible, since fewer people just meant less work done.

So when the headcount doesn’t come, you get creative with what you’ve got. You bring in a contractor or an agency to cover the gap. Or you take the program dollars you can’t turn into a hire and try to spend them where they’ll actually move the work forward. Or you buy a tool and hope it does the job the extra person would have. Either way, you make cuts, you make concessions, and you find a way to deliver, but none of it solves the underlying problem. A tool dropped on top of a broken process just makes the broken process faster.

The real work has always been balancing those four things, headcount, program spend, the contractors, and the tools underneath, to improve the value you deliver and not just the volume you push out. What’s changed is the reflex. Because customer marketing has mostly been a tactical function, we learned to assume that more workload means more headcount. That isn’t automatically true anymore.

A person plus a set of narrow agents can now cover a pillar that used to take a full-time hire, and that tips the whole balance. The org chart turns into an operating model: one practitioner at the center, running a set of agents, each pointed at a pillar, with a governance layer that keeps the whole thing honest. And because the work lives in that system instead of getting rebuilt by hand each cycle, the next round starts where the last one left off, and not from a blank page. That’s the loop we didn’t create when I first joined Marketo.

I want to be careful here, because there’s a version of this that sounds like a layoff pitch, and it isn’t one. The agents aren’t there to replace the customer marketer. They take the part of the week that was never customer marketing in the first place: the account research, the CRM digging, the spreadsheet aggregation, the fourth rewrite of the same account summary. That’s the work that pulls you away from customers. Hand it off, and the person gets their week back for the part only a person can do, which is the relationship building.

Five things make the model work, and none of them are necessarily exotic. The context each agent reads lives in plain files you own, one canonical home for each thing, so every agent works from the same truth. Each agent is narrow, a specialist for one pillar instead of one bot trying to do all seven. Every agent checks its work against a real source instead of guessing, which is what earns the trust of the people who’ll ask how you know. Anything that goes to a customer is drafted by the agent and sent by a human, every time, because that’s where judgment and reputation live. And nothing gets built in the abstract: each agent starts from one real task you already repeat every week, and gets better by running for real and fixing what breaks. You start with two, not a hundred.

If you want to feel this in your own week, don’t start by building anything. Write down what pulls you away from your customers, the research, the CRM digging, the spreadsheets, the summaries you redo every week, and point your first agent at the top of that list. And fix your data before you point anything at it, because an agent reading a messy source just makes it worse.

None of this is a someday idea. It’s the argument I’m walking a room through next week at the Customer Marketing Summit in SF: the operating model, the org chart, and a working version of it running on my own operations. This issue is the short version.

This week’s free starter: the pillar-to-agent map

A one-page worksheet. List your seven pillars down one side, and for each, name the single repeated task an agent could take off your plate first. That map is where you start, before you build anything, and it’s the whole “start with two, not a hundred” idea in the smallest form I could make it.

Access the Pillar-to-Agent Map

For paid subscribers

The map above isn’t theoretical for you, because your kit already has the pillar archetypes, agents mapped to each of the seven pillars. So this week, pick the two pillars dragging you most, open the matching agents in your Blueprint, and stand them up on one real task each. The task briefs are written, so you’re not starting from a blank page. By Friday you’ve got two pillars running on the operating model instead of on your evenings.

Access My Complete CLG Blueprint

— Kevin

P.S. If you’re the one or two people covering all seven pillars, forward this to the leader who keeps asking for more without more headcount. It’s a better conversation to have than the one where you quietly burn out proving you can do it all.

Welcome back to The Customer Continuum. Issue #36.

A few years ago I brought a handful of our most active community members into a room with the new head of our support team. I set it up for two reasons. One was to build a little trust between these customers and a leader who was new to them, and between that leader and me. The other was to get at a harder question, which was how we bring these two groups together to solve something that hadn’t really been anyone’s job.

These are customers who go out of their way for us. They answer other people’s questions in our community, they jump into Reddit threads and other forums, and a lot of what they sort out never becomes a support ticket at all, because they’ve handled it for someone before it ever reaches us. What came out in that room was that the people doing the most for us weren’t always getting that same care back. They’d help everyone else and then wait longer than they should have on their own issues.

None of that was in a report. It took putting the right people in one room to even see it, and because the new support leader was sitting there with the customer, we could start solving it together immediately.

That afternoon stuck with me. What we needed to see lived in how those customers were spending their time for us, and in what they weren’t getting back, and it only surfaced because we brought the right people together to look. A dashboard or a survey would never have shown us any of it. That’s the idea behind what I built this week, something that reads the signal customers are already giving off and puts it in front of the person who can act on it.

Last week I built a one-time read of customer signal, an agent that looks across the places customers actually talk and hands you back the themes. This week I make it run every week, so it starts to show me what’s changing from one week to the next.

Voice of Customer, or VoC, is just the practice of capturing what your customers are telling you and turning it into something the business acts on. For most teams that’s meant surveys. You send the survey, you wait, and you read the numbers and share the report to your executive team. Often times, the data is already stale.

A single read tells you how your customers feel this week. Run it every week and you start to see movement and patterns: which themes are climbing, which ones just showed up, which ones went quiet. One read tells you where a theme sits today, and the change from one week to the next tells you where it’s heading.

That matters more than it sounds, because the signal that costs you the most money rarely arrives all at once. A theme shows up faintly one week, a little louder the next, and by the time it’s loud enough for a single read to flag it, the account it’s attached to has often already made up its mind. A standing pulse catches it while it’s still building. And catching it is only half the job, because a theme with nobody’s name on it is just an observation. So the pulse does two things every week: it names what’s changed, and it points each theme at the person who can actually act on it.

What I built

I didn’t build anything new this week, and that’s the point. Last week’s issue put two files in the kit, the multi-source signal agent that reads across calls, tickets, community, and usage, and the routing layer that assigns an owner to each theme. This week I wired them into a cadence. Same agent, same rules, run on a weekly schedule with one job added: compare this week’s themes to last week’s and report what moved.

The report comes back in three buckets. New themes, ones that weren’t there last week. Escalating themes, ones getting louder. And themes that went quiet. That third bucket, the quiet one, needed a rule of its own that the one-time read never did.

A theme going quiet doesn’t mean the problem is solved. When a problem stops showing up in the signal, it might mean you fixed it, or it might mean the customer stopped bothering to tell you because they’ve already decided to leave. From the outside those two look identical, and the diff can’t tell them apart. So the pulse is never allowed to close a theme on its own. It reports the quiet as a status to go confirm, and it hands it to someone to check.

The two rules from last week carry over unchanged, because they matter more in a weekly cadence than they did in a one-time read. The first is that the agent never invents a theme it can’t cite. If the evidence is one loud comment, it says so and marks it as something to watch rather than dressing it up as a trend. The second is the absence rule, the one I care about most. The agent can report that there’s no signal about something, but it’s never allowed to read silence as satisfaction, because a quiet account isn’t a happy one. Run those rules once and they keep you honest for a week. Run them every week and they’re the only thing that keeps a standing pulse from quietly turning into false comfort.

The honest run

So I ran it, the way I run everything in this newsletter, cold and then torn apart by a second agent whose only job is to find where the first one was wrong. I built two weeks of signal in a fictional set of accounts, deliberately messy the way real signal is, and treated the first week as the baseline, the “last week” the pulse compares against. Then I fed it a second week and asked for the diff, without telling it what I thought it would find.

It caught exactly what I’d hoped a pulse would. In week one, onboarding was a middling theme. It was in the data, a couple of accounts grumbling that ramp-up took too long, but it didn’t stand out, and a single read that week rated it moderate and moved on. In week two it climbed hard. A customer froze an expansion because the last rollout had taken a full quarter to get productive. A second account paused adding seats for the same reason. An account churned and named a slow start as the reason they stopped seeing value. Onboarding tickets jumped by about a third. Put week two next to week one and the shape is obvious, a revenue problem building in real time, and the diff pushed it to the top of the pulse. A single snapshot taken the same week would have shown it sitting about where it had been.

It also flagged something completely new, and it read it correctly, which honestly surprised me. A handful of customers asked for a feature the product already has, a way to schedule reports. One of them wrote their own script to do a thing that ships in the app today. A community post asking for it drew a top reply pointing out it already exists. The agent didn’t file that as a missing feature. It called it what it was, a gap in what customers know rather than a gap in what we’ve built, which is a completely different fix with a different owner.

And then the theme that had been the loudest the week before went quiet. Last week, the strongest signal in the set by a distance was customers struggling to prove the product’s value to their own finance leaders. This week, nothing. No new calls, no new posts, the big community thread from the week before just sitting there with no new replies. This is where that rule provided the most value. The agent refused to call it resolved. It reported the theme as gone quiet and unconfirmed, and it went looking for what was underneath the silence, which turned out to be two things worth worrying about. One of those finance-buyer accounts had gone dark after we sent them a value summary. Another had a senior person skip their last check-in, with a renewal about two months out. That silence is exactly what the absence rule exists to distrust, because it looks identical to a problem that got solved.

So far that’s the run going right. It also fumbled, in two ways.

The first is routing. It got the onboarding theme dead right and then handed it to the wrong mix of people. It assigned it to customer success and sales, which makes sense for the churn and the frozen expansions, and it left product off entirely, even though the one concrete, fixable thing in the whole onboarding story was a setup checklist that had gone out of date, which is product’s to fix. The theme needed three owners and got two. What makes it worse is that the agent knew. In its own self-check it admitted it had probably under-assigned product, and then it shipped the routing anyway. Handing the theme to two teams that can’t touch the actual fix means the root cause just sits there while everyone agrees it’s important.

The second fumble is quieter. When that ROI theme went silent, the agent correctly refused to call it resolved, but it still ranked it near the bottom. Its logic was mechanical: fewer mentions this week means weaker signal, and weaker signal sorts low. So the single most dangerous account in the whole pulse, a strong customer who went dark two months before a renewal, ended up filed under a label that reads like “minor, keep an eye on it,” sitting right next to a genuinely minor item. A reader skimming the pulse would go straight to the onboarding theme at the top, the one they already knew about, and walk right past the account most likely to churn, sitting near the bottom under a low-priority label.

That last one is the failure I didn’t expect. The diff got good at spotting what’s moving, but the way it ranked and labeled things didn’t keep up. A theme can now be quiet and dangerous at the same time, and the labels I carried over from a one-time read just don’t have a word for that. If I’d nailed everything, I’d distrust the run. What held the whole way through was the absence rule. Every week it flagged that the survey feed still wasn’t wired instead of treating the silence as good news, and it refused to read a quiet customer as a happy one. That’s the discipline that makes a weekly pulse safe to run, instead of something that just sounds a little more sure of itself each week while it’s really still guessing.

What this means for your team

You don’t need to build any of this to use what it’s telling you. The move underneath it works with a notebook.

Take the two or three themes you’d say are live with your customers right now and write them down. Next week, write them down again. The gap between the two lists is the entire point. A theme that was quiet and is now everywhere is escalating, and it’s telling you something a single glance this week can’t. A theme that’s brand new is worth a second look before it gets loud. And a theme that was loud and went silent is the one to handle most carefully, because you can’t tell from the quiet alone whether you fixed it or lost them, and treating it as fixed is how a save turns into a surprise.

Then give each one an owner, a real person who can act, not a function in the abstract. And hold the two rules while you do it. If a theme rests on one loud voice, mark it to watch instead of calling it a trend. And if an account goes quiet, don’t assume that’s good news.

What’s next

Next week I’m taking the pulse off my desk and putting it on a real schedule, wiring it to run on its own and drop each week’s diff where the owners will actually see it, instead of depending on me to remember.

This week’s free starter: the week-over-week diff

The free starter is the diff on its own, pulled out of the full pulse so you can run it by hand. You give it two lists, the themes you saw last week and the themes you’re seeing this week, and it hands back what’s new, what’s escalating, and what went quiet, with the “quiet is not resolved” rule built in so it can’t quietly close a theme on you.

You’ll see it right away. Put last week’s list next to this week’s, and a theme you’d have sworn was steady shows up climbing. A single read can’t show you that.

Drop it in a fresh Claude Project, paste your two lists, and read what it says moved.

Week-over-Week VoC Read (Free Starter)

For paid subscribers

The free starter diffs two lists you typed. Your kit runs the whole thing on your own signal, every week, without you steering it.

The honest part, and the whole pitch this week: you already own the pieces. Last week’s issue put two files in your Copilot, the multi-source signal agent that reads across your calls, tickets, community, and usage, and the routing layer that assigns each theme an owner. I didn’t ship you new files this week. I’m handing you the task briefs to stand up what you already have as a running weekly pulse, so by Friday it’s a habit instead of a thing you did once.

The promise is concrete. By the end of this week you have a weekly VoC pulse running on your actual signal, every theme routed to an owner, and a week-over-week diff in your hand that tells you what’s climbing and what’s gone quiet. You run it once to set the baseline, and after that it’s a standing rhythm on the signal you already generate every week.

This week’s build, task briefs written so you don’t have to design it:

→ Stand up the multi-source agent on your own four feeds and run it once to set this week’s baseline:

Access the VoC Signal Agent

→ Turn on the routing plus weekly cadence, and schedule next week’s diff against this week’s baseline:

Access VoC Theme-to-Owner Cadence

Next week, putting the pulse on a real schedule so it runs without you.

— Kevin

P.S. If reading this made you think of one account that’s gone quiet on you lately, don’t file that quiet as good news. Forward this to one customer marketer who’s about to close a theme they should be confirming, and then go check on that account. That’s how this grows.

Welcome back to The Customer Continuum. Issue #35.

A few years ago we nearly lost a large account, and the only reason we caught it in time was that someone on my community team happened to be reading the room.

We ran a customer community and a formal survey program, and they were two completely separate workstreams. Different teams, different tools, different calendars. The survey team ran the relationship survey and the pulses on their own cycle, and the community team did what community teams do, working through the discussions as they surfaced. One of those community managers, going through the day’s threads, noticed a conversation from an account we cared about starting to turn. Nothing had blown up yet. It was early, just a tone shifting in a thread, the kind of thing that’s easy to scroll past, and she flagged it anyway because it read like something that would fester if it sat.

She was right. Underneath the thread was a customer in a bad stretch, dealing with an outage on their side that had spilled into the service they were getting from us, and quietly starting to wonder out loud whether they’d be better off somewhere else. Left alone in the queue, that’s a churn note we’d have read a quarter later. Caught early, while it was still just a concern surfacing, it was something we could get ahead of.

Our formal read on this same customer was blind to all of it. The relationship survey, the transactional pulses, the Net Promoter Score program, the survey that asks how likely a customer is to recommend you on a zero-to-ten scale, none of it was going to say a word until its next scheduled cycle, a quarter out. And the survey team and the community team weren’t even looking at the same account in the same week, because they were built as separate workstreams. The signal that mattered was sitting in one team’s queue while the instrument that was supposed to measure how the customer felt was pointed somewhere else entirely.

So we did the thing the survey couldn’t. We took the early signal to our product leadership, stood up a triage with the account team around the outage the customer was actually living through, and moved on it in days instead of at the pace of a survey cycle. The account stayed. We kept a six-figure relationship that had started weighing its options, and the customer went from quietly doubting us to working with us on the fix.

What stayed with me was how fragile the catch was. It came down to one person on the community team reading a thread the survey team would never see, while the survey we trusted to tell us how customers felt sat silent for another three months. The signal was there the whole time. Our instrument just wasn’t pointed at it, and the team that could see it wasn’t wired to the team that ran the numbers.

The old Voice of Customer, and the one that’s replacing it

Voice of Customer, or VoC, is the practice of capturing what customers tell you and turning it into something the business acts on. For most teams it’s meant surveys, sent on a schedule and read a quarter later: the Net Promoter Score survey, the customer satisfaction survey (CSAT, the “how did we do” rating after an interaction), and the product satisfaction survey (PSAT, the same idea aimed at the product). You send them, you wait, and you build your quarter around the number that comes back.

The trouble starts with how few people answer anymore. Survey fatigue is real, and the numbers behind it are hard to argue with. The email response rates that carry most NPS and CSAT programs have slid from the low-to-mid twenties a few years ago into the low teens today, and one major feedback platform measured its own rates falling by more than a quarter in just four years. The average person now fields three to five feedback requests a week, so your survey lands as one more email in a pile of receipts and delivery pings, and most of them never get opened.

The customers who do answer are a skewed sample, and the bias matters even more than the shrinking response rate. The people who fill out a survey tend to be the ones at the extremes, delighted or furious, while the quiet majority in the middle skips it. So the number you build your quarter around comes from a small, self-selected slice of your base, often a tenth of it or less, weighted toward the loudest voices in the room. And the account that’s quietly falling apart is usually the first to stop responding, so the customer you most need to hear from is the exact one your main instrument goes deaf to.

Even the signal you do capture sits in pieces. The community team sees the thread, support sees the ticket, the product team sees the usage, and the survey team sees the score, but no one sees the whole customer, because the picture is scattered across teams that never compare notes until it’s a churn note. And that’s only the signal inside your own walls. Your customers are also talking about you in the channels you don’t own, on Reddit, in G2 and Capterra reviews, in public threads where you’re nowhere near the conversation, and that’s often where they’re most honest, precisely because you’re not in it. Most teams have nobody assigned to read any of it. What’s changed is that we can finally read the customer across all of it at once, the calls, tickets, threads, and usage you generate, and the public signal you don’t. The survey doesn’t go away. It just becomes one input among many instead of the only one you trust, and the job now is to be deliberate about assembling the whole picture rather than cherry-picking a slice of it.

The new VoC is about closing that gap. It reads the customer across every one of those sources in close to real time and pulls the scattered pieces into one picture instead of four disconnected ones. It doesn’t retire the survey. It stops treating a quarterly score as the whole truth, when a fuller and faster truth is already sitting in the calls and tickets and threads you generate every day. Community especially gets written off as noise, when it’s often the earliest and most honest signal you have, the place a customer says the thing they’d never put on a five-point scale.

What I built

The Voice-of-Customer signal agent is straightforward to describe. You point it at raw signal, one source or several, and it reads across all of it at once and hands back the top themes. Each theme is stated plainly, each has a suggested owner, and each carries the specific evidence underneath it, the actual ticket or call or post or usage number. The point isn’t any single source but that the agent reads them together, so a concern that shows up faintly in a call and loudly in a community thread gets connected into one theme instead of missed by four separate teams.

The four sources it reads are the four places customers actually talk: the call transcripts a tool like Granola captures off your meetings, the support tickets and churn notes where the friction and the losses live, the community activity in your forum and public threads, and the product-usage signals that show you what people do instead of what they say. A survey is a fifth source that still works it’s usually lagging data.

I gave it two rules that matter more than anything else it does, because without them a summarizer like this becomes a confident liar.

The first rule is that it never invents a theme it can’t cite. Every theme has to point at real signal. If the evidence is thin, it has to say so and mark the theme as weak and worth watching, rather than dressing up one comment as a trend. An agent that turns a single loud complaint into “customers are frustrated with X” is worse than no agent, because now you’re acting on a number you made up.

The second rule is the one I learned building my chief-of-staff agent earlier this year, and it’s the one I care about most. It’s the absence rule. “There’s no signal about this in the support tickets” is a fact the agent can observe and report. “The customer is happy” is a claim that no single source, and no amount of silence, can support. A customer who’s gone quiet might be content, or might have stopped bothering to tell you anything because they’ve already decided to leave, and you can’t tell which from the quiet alone. The agent has to report what it can see and name what it can’t, and it’s never allowed to read silence as satisfaction.

The honest run

I ran it cold against one sample dataset spanning all four sources, deliberately messy the way real signal is, some of it rich and some of it nearly empty, and I pointed the agent at it without telling it what to find. Then I had a second agent tear the first one’s work apart, which is the same operator-plus-checker discipline this whole newsletter is built on. Here’s what came back.

The theme it nailed was the one a survey would have missed entirely. Reading across all four sources, it surfaced that customers couldn’t prove the product’s value to their own finance leaders, and it built that theme out of four completely separate pieces of signal that never mention each other. In one call a chief financial officer said nobody had ever shown him a number, and in another a customer said the dashboards were fine for operators but couldn’t go in front of their CFO. A community post asking how anyone shows return on investment to leadership drew a top reply calling it the number-one blocker to expanding a contract, and a churn note from an account that had already walked cited that they never proved value to their finance team. That’s four voices in four sources landing on one theme, and not one of them would have shown up as a bad survey score, because the people filling out the survey were the happy operators rather than the finance buyer quietly deciding the renewal wasn’t worth defending. It’s the exact save I lived years ago, watched a machine find on its own, from signal a survey structurally cannot reach.

Then it over-read the data in two different ways worth showing you, because they’re how this kind of agent fails in the wild.

The first was a single incident dressed up as a trend. There was a real and strong theme in the data about a feature with no safe way to preview what it would do before you ran it, and the agent caught it correctly, backed by rising abandonment in the usage numbers and a stack of tickets. But it went one step too far and called it a pattern of the feature “causing data loss,” when the actual data-loss evidence was one escalated ticket and one anecdote in a forum. One incident is something to watch and confirm, and if you carry it to your product team as a trend you’ll spend a sprint on the wrong severity.

The second was reading behavior as a complaint. The usage data showed people opening the reporting dashboards for forty seconds and exporting to a spreadsheet in most sessions, and the agent decided that proved the dashboards were too shallow to use. Maybe it does. Or maybe those people are perfectly happy pulling data into their own decks because that’s how they work. The same numbers fit a frustrated customer and a content one equally well, and the agent picked the frustrated story because it had one loud account in the mix, when the honest read was to hold it as a question until a second account confirmed it.

The last thing it got wrong is the one that stuck with me. The biggest money signal in the whole set was onboarding dragging on, showing up again and again as the reason accounts wouldn’t expand. Two lost expansions, a sales note that it was the second that month, a churn that named it, a support ticket about docs that didn’t match the product. That’s real revenue walking out the door. The agent found the signal, rated it middling, and nearly buried it. It played it safe because the clearest quote came from an internal sales call instead of the customer directly, and one of those deals had a pricing issue mixed in. So the same caution that keeps the agent from crying wolf talked it out of the one signal most tied to lost money.

The scorecard came out like this. One theme it nailed, the kind a survey would never have caught. Two it over-read, one by turning a single incident into a trend, the other by reading normal behavior as a complaint. And one big revenue theme it nearly missed by being too careful. If it had gotten all four right, I wouldn’t trust the run. What held the whole way through was the absence rule. It wouldn’t call flat usage a sign of health, and it wouldn’t treat a missing survey feed as proof anyone was happy. When a customer exec opened with “you’re doing fine,” it flagged that as a polite line in front of a budget worry, not a happy account. Every single time, it told me what it could see and what it couldn’t. That’s the reason I’d trust it near a real list of accounts, instead of a tool that sounds reassuring while it’s really guessing.

What this means for your team

You don’t need to build this agent to use what it’s telling you. The move underneath it is available to you on Monday with nothing but the sources you already have.

Your customers are telling you how they feel right now, this week, in places you’re not systematically reading, whether that’s the last three QBR transcripts, the support tickets that keep reopening on the same account, the one community thread that’s gotten a little too honest, or the feature whose usage quietly dropped off a cliff for your biggest customer. Every one of those is Voice of Customer, and every one of them is faster than the survey you’re planning to send next quarter. The survey is still worth sending. It’s just the slowest witness you have, and most teams treat it as the only one.

So this week, take one account you’re even slightly worried about and read it the way the agent does. Pull the signal from more than one source, a call, a ticket, a thread, a usage number, and write down the two or three themes that actually repeat across them. Give each theme an owner who can act on it, because a theme without an owner is just an observation. And hold the agent’s discipline while you do it. If a theme rests on one loud comment, mark it as something to watch instead of a trend. And never let yourself read a quiet account as a happy one, because the account I almost lost went quiet in our surveys right before it surfaced in our community.

What’s next

Next week I’m wiring this from a one-time read into a standing pulse, the version that runs on a weekly cadence across all four sources and routes each theme to the owner who can act on it, so it works like a real-time VoC read instead of a thing you run once and forget.

This week’s free starter: the single-source signal read

The free starter is the agent scoped down to one source. Point it at a single export, your last batch of call transcripts or a dump of recent support tickets, and it hands you the top three themes with an owner on each, holding both rules, no invented themes and no reading silence as health. The quick win is watching a theme surface that your survey would have missed, from signal you already had sitting in a folder.

Drop it in a fresh Claude Project, paste in one export, and read the three themes it gives you next to what your last survey told you about that same account. The gap between the two is the entire point of this issue.

Access the VoC Signal Agent

For paid subscribers

The free starter reads one source. Your kit reads all four and runs on a schedule.

The full Voice-of-Customer signal agent is wired across every source at once, call transcripts, tickets and churn notes, community activity, and product-usage signals, so a theme that shows up faintly in one place and loudly in another gets connected instead of missed. On top of the read sits the theme-to-owner routing, so every theme comes out already pointed at CS, product, marketing, or the exec who can move on it, and a weekly cadence so it runs as a standing VoC pulse rather than a one-off you have to remember to trigger.

The promise is concrete. By Friday you have a real-time Voice-of-Customer read running on your own signals, the calls and tickets and threads you already generate every week, instead of a survey you send next quarter and interpret the quarter after that. Not a framework for building one. The working agent, reading your actual sources.

This week’s build, task briefs already written so you don’t freeze in front of the whole library:

Access the Paid VoC Signal Agent

Access the Theme to Owner Routing Agent

Access the Full 135 CLG Blueprint

Next week, the standing weekly pulse build.

— Kevin

P.S. If this made you think of one account you’ve been reading through a survey when the real story was sitting in a call transcript, forward it to one customer marketer who’s still waiting on next quarter’s Net Promoter Score for news a support ticket could have given them weeks ago. That’s how this grows.

Welcome back to The Customer Continuum. Issue #34.

The last few issues, I built the champion outreach system. This week’s a different animal.

A few years ago I pitched a program my leadership thought was a waste of money. We wanted to fly a handful of our most important customers to an executive forum in Sardinia, Italy. Yes, Sardinia. The CRO looked at me like I’d tried to expense a holiday, and I understood the reaction, because on a spreadsheet that’s exactly what it is, a line item with a beach behind it.

The idea underneath it was bigger than one trip. Almost everyone in our world was a practitioner, the people who live in the product day to day that we had a deep relationship with. The executive rooms above them were a complete blind spot. We had no real relationships with the people who actually sign the renewal, and no executive community to speak of. So we partnered with a third party organization that ran exactly these kinds of gatherings and curated ours together. I pitched it as a pilot on purpose, a small experiment to find out whether investing in executive engagement would pay off at all, because I couldn’t promise anyone that it would. The skepticism was fair. I was asking for real money on a bet nobody had proof for yet.

What the spreadsheet couldn’t tell anyone was that one of the accounts on that invite list was quietly falling apart. Not loudly, and not in any way that shows up on a renewal forecast, but the signals were everywhere if you were paying attention: support tickets that kept reopening, a tone in the reviews that had drifted from partner back to customer, the slow and specific feeling of a company that had stopped feeling valued. We’d been reading those signals for months. So the forum was doing two jobs at once. It was the test of a new idea, and it was a move on that specific account, a way to stand up an executive sponsorship program around them, put our leadership in a room with theirs, and answer what the customer had been telling us for a year.

That trip is where it turned. Their execs came. Ours showed up for them, and I mean actually showed up, not a handshake and a photo. The account that was halfway out the door stayed, we saved a deal worth more than $500,000, and the customer went from writing us off to helping us rebuild the relationship from their side. And the pilot did exactly what it was intended to do. It gave us a positive outcome: YES, keep building out the program, which is how a one-time experiment turns into a program you get funded to keep running.

What stuck with me came a quarter later, watching that story get compressed in a business review. It became a single line, that the Sardinia trip saved the account, and everything underneath it disappeared: the months of reading the signals, the bet on executive engagement that got us in the room at all, my team catching the drift early enough to act, all of it flattened into one slide about a nice trip to Italy. The save was real, but I hadn’t learned yet that saving the account and getting credit for it are two different skills, and I’d only built one of them.

I’ve sat in that hot seat more times than I’d like, defending a budget and explaining what my team actually did to people who’d already decided what they thought it did. What finally clicked for me is that it was never about the tactics. I could recite the program list in my sleep and it wouldn’t move a single person across the table. What I had to get good at was saying the same work in a different language depending on who was sitting there, the CRO, the CFO, the CEO, the CMO. That took me years to learn. So this week I tried to build it into an agent.

This is really an attribution problem before it’s a marketing problem. The work that saves or wins a deal happens upstream and quiet, the trust you build over a year, the champion who’ll go to bat for you, the signal you caught before anyone else was looking. The moment a deal closes is loud, and it belongs to whoever’s in the room when it happens. Your quiet year of work can be the whole reason that deal was winnable and still not be the story anyone tells. The version that sticks is the one spoken in the language of the number leadership already watches.

What I built

The Decoder Ring is simple to describe. You give it two things, what you did, whether that’s a metric or a program or a deal you helped save, and who you’re about to sit across from. It hands back your value in that leader’s language, pointed at the outcome they actually own, with the tactics stripped out.

Underneath, it runs the three moves I use when I walk into one of these rooms myself. Open on the goal they already care about instead of on my function. Restate what I did as a number they use to run the business. Then stretch the frame out to the second-order thing they’re measured on. It’s built on a belief I’ve earned the hard way, that you don’t win this in a meeting. You win it over cycles, by speaking their language and building the cross-functional trust in parallel, until one day someone else is telling your story for you in a room you’re not even in.

I gave it two rules that matter more than the moves. It won’t hand you one reframe to spray at everyone, because the CFO and the CRO don’t own the same outcome, so it makes you pick the room before it writes a word. And when it doesn’t have the real number, it won’t invent one, it tells you what’s missing and asks for it. A confident reframe you can’t back up is worse than saying nothing, and I’ve watched that go badly in real time in front of a CFO.

None of this is a fresh rebuild for every leader. The C-suite wants the same handful of things almost universally: be brief, be executive-ready, and speak in outcomes instead of a list of tactics. Get that right and you’re most of the way there. The per-seat tuning, the CFO’s cost-of-acquisition lens or the CRO’s expansion lens, is the last mile on top of a translation that’s already doing the universal job, turning what you did into what it meant for the business.

The honest run

I ran it cold against a real number: customer advocacy influenced $21.1M in ARR in my last full year running the practice. I pointed it at three leaders, one at a time. Two of the decodes I could have used as they were. It’s the third that stuck with me, because it didn’t just miss but showed me where the whole idea still breaks.

The CFO version I’d have walked into a boardroom with. It opened where finance always starts, on cost of acquisition, and recast the $21.1M as revenue that closed or renewed with an advocacy touch on it, a reference call, a peer review, a champion intro, all booked without paying to source it cold. Then it did the thing I’d told it to do instead of bluffing, and admitted it couldn’t give me the real efficiency number without my blended CAC and my attribution window, so it asked for both. It even landed where I would have, on the move from per-seat to consumption pricing, where every renewal has to re-earn its keep and advocacy is the cheapest proof you’ve got.

The CRO version started just as strong, leading with the give-and-get I’ve always run with revenue: I build the reference desk and the user groups that shorten your cycles and let customers do some of the selling, and you send me your champions and the deal context. Then it slipped. It called the $21.1M “mostly expansion” as if it knew the split, and it doesn’t, because I never gave it one. That’s exactly how our function loses the room, claiming the shape of a number you can’t break down the second someone asks.

Then the customer success leader, where the whole thing fell over. A CS leader already believes in you. They own the retention number, they can see the programs working, they’re a partner, not a skeptic you have to win. So everything the agent is good at backfired. Co-opting someone who’s already on your side does nothing, and anchoring in revenue just tells them what they already know. What came back was accurate and completely useless: “your customers are healthy, adopting, and expanding, which supports your net revenue retention.” It handed a partner a compliment when what they needed was a play, which account, which signal, what to do Monday morning. The Decoder Ring knows how to win over someone who doubts you, and it has no idea yet what to do with someone who’s already in your corner.

Two clean, one flat, and the flat one taught me more than the wins. I’ll explain more below.

What this means for your team

The takeaway isn’t “go build a decoder agent.” You don’t need one.

It’s this. Every piece of work you own is already three or four different sentences, depending on who’s reading it. And your good work goes uncredited in meetings less because it was weak and more because you walked in speaking marketing to a room that only hears revenue. So before your next big conversation, take your single strongest result and write it three ways: a cost-of-acquisition line for the finance brain, an expansion-and-retention line for the revenue brain, and an operational signal for the success partner who’s already rooting for you. I’d bet the version you’d have led with isn’t the one that lands.

Which raises the obvious question, the one I asked myself: what happens when they’re all in the same room and you’re the one presenting? You don’t run three translations out loud. You lead with the shared business outcome, briefly and in plain terms, because that’s what every seat at that table wants anyway. The tuned versions become your backup, the answer you reach for when the CFO or the CRO leans in with their specific question. The universal move carries the room, and the tuning is there for the follow-up.

The translation is only half of it. The other half is planting seeds. You make the case yourself, in the language each leader trusts, enough times and over enough months that eventually someone else starts making it for you in a room you’re not even in. When you’re the one saying it, leadership hears one more person defending their own budget. When the CRO says it for you, they take it as fact. You earn that by doing the translation over and over, and by building the cross-functional relationships long before a renewal is ever at risk, so that when the account gets saved, someone who isn’t you tells the room your team saw it coming. It’s slower than making a good slide, and it matters more.

And steal the agent’s one discipline while you’re at it. When you don’t have the real baseline, say so and ask for it, instead of reaching for a number you can’t defend when someone leans in. “I don’t have the CAC baseline yet, get me that and I’ll show you the efficiency” beats a confident guess that falls apart the one time it counts.

What’s next

Next week I’m building the ally mode this run exposed, the version that equips a partner who’s already on your side with a play instead of a pitch, which is exactly the move the customer success decode was missing.

This week’s free starter: the C-suite decode

The free starter is a scoped slice of the Decoder Ring, the C-suite decode on its own. Feed it one result and it hands you the CFO or CEO version, anchored in cost of acquisition, with that same rule that it asks for a missing number instead of inventing one. It’s the fastest way I know to see how much you leave on the table every time you carry marketing language into a revenue room.

Drop it in a fresh Claude Project, paste in your strongest result, and read the decode next to how you’d have said it yourself. That gap is the point.

Access the C-Suite Decoder Ring

For paid subscribers

The free starter decodes for one audience. Your kit decodes for all of them.

Read more

Welcome back to The Customer Continuum. Issue #33, the finale of the champion series.

We’ve done the hard part twice now. One build found the champions: an agent that reads your customer base and points to who’s worth pursuing. The next reached them, with an invitation built around what that specific person had actually done. Getting champions in is the front of the work, and it’s the part that gets all the attention. This issue is about the part nobody wants to talk about, which is what happens in month four, when the launch glow is gone and the calendar has moved on and the program is either running or it isn’t.

I learned this the hard way with Adobe Analytics.

When I stood up the champion program there, I didn’t sit in a conference room and guess at what would keep people engaged. I called them. I talked to fifteen or twenty of the customers our product and customer success teams had nominated, and I asked them straight out what they cared about, what their hot buttons were, what made them tick. What surprised me was how little separated them from the Marketo customers I’d spent years with.

These were data people. Nerds in the best sense, the ones who build the workflows and the analytics architectures, who’ll argue Adobe Analytics versus Google Analytics for an hour, who geek out over a clean data model and a well-drawn system diagram. Their daily grind was living in that data and then translating it, turning it into something a marketing team or a C-suite exec would actually act on. I figured they’d want something completely different from the marketing-automation crowd. I found out they wanted the same three things.

Peer networking came first. They wanted to be in a room with people who did the same job and spoke the same language, who understood what it’s like to be the data person everyone leans on. Second was recognition. Their name in lights, the kind of career exposure that shows up in a performance review, which is why we built feeders straight into the Experience Maker Awards and the user groups. The clearest example I’ve got is Jess Kao. She leaned all the way into the awards and leading a user group, and it reshaped her career, a run of recognition and a kind of visibility she’s carried ever since. The third was access. A shorter path to the people building the product, a seat in the betas, a hand in shaping the education courses. They wanted in on the actual work.

Peer networking, recognition, access. Give a champion those three on a schedule and they stay. Skip a quarter and they drift, and you usually don’t notice until the renewal date has already slipped past.

That became the architecture, and it outlasted the org chart. What started as Advocate Nation still runs today inside the broader Adobe Experience Cloud portfolio, the same shape I built back then: the Experience Maker Awards, the reference program and customer stories, the user groups, the voice-of-customer work, the adoption and retention plays, the champion program itself, the webinars. It didn’t quietly wind down the way most of these do. The reason is honestly kind of boring. Somebody just kept running the cadence long after the launch was over, which is the whole job this issue is about.

What I built

So this week I built the engine that runs the keeping. It’s called the Champion Program Agent, and it does the unglamorous work that starts the day after a champion says yes.

Quick context if you’re new here. A champion program is a small, invitation-only group of your most influential customers, the ones who’ll advocate, speak, refer, and mentor other customers, in exchange for recognition and real access. The first two builds got them in the door. This one keeps them there, and it runs on three jobs.

Tiers first. The agent sorts champions into levels, Community Champion at entry, Product Champion in the middle, Executive Champion at the top, and it re-checks the advocacy score every month to catch who’s earned a bump up and who’s gone quiet and slid down. Nobody keeps this up by hand past the first quarter, which is exactly why the agent does it.

Then benefits. When someone moves up a tier, the agent fires off that tier’s benefits and posts the recognition, so the thing you promised them at signup actually lands in their inbox instead of dying in a spreadsheet you meant to get to when you had 15 min.

Then renewal: every champion has an anniversary, a year from the day they joined, and the agent sends the re-invite sixty days ahead of it and confirms they’re still eligible. That’s all renewal means here, the yearly ask to stay another year. Miss it enough times and the cohort lapses because of neglect and not providing the same level of attention as when you started the program.

The honest run

Then I ran it. Cold, against the seven champions who accepted in last week’s build, with nothing else attached, no tiers, no benefits on record, no join dates. These are a test cohort, not a live roster. I wanted to see the engine’s honest floor, what it can do on day one before anyone’s had time to clean up the data.

It did nothing, and that turned out to be the best outcome.

It assigned zero tiers, out of seven. The three levels exist on paper, Community, Product, Executive, but the agent places a champion by reading an advocacy score against each tier’s criteria, and this cohort showed up with neither. No scores. No thresholds written down. It could tell from the recruiting notes that a couple of these folks had deep track records and others were lighter, but a hunch isn’t a tier, so it wouldn’t stamp one on. It flagged all seven and told me what it needed first.

It couldn’t confirm a single benefit. With no tier there’s no benefit set to check against, but the real gap was deeper. Nothing on record said what any champion had been promised at signup, or whether one thing had ever been delivered. So all seven came back the same, unconfirmable, because the program had never kept a ledger of what it owed these people.

It couldn’t schedule one renewal. The renewal step fires sixty days before a champion’s anniversary, and not one of the seven had an anniversary on file, so there was no way to know when to re-invite anyone. Two got flagged a second time, the agency and partner champions whose eligibility nobody ever resolved when they were first reached, because the agent won’t queue a renewal for someone the program hasn’t confirmed is even allowed to be a member.

Here’s what all of it had in common. The program was built to bring champions in and never keep track of their engagement and records. No join dates, so renewal has nothing to run on. No written criteria, so the tiers are just names. No benefit ledger, so no way to prove you kept your word. No score feeding the monthly check that’s supposed to catch someone going quiet. Every one of those is something that should have been created the day a champion signed, and wasn’t.

I’ll be honest, I expected it to at least limp through a few of these. It didn’t, and the zero is the lesson. The engine ran clean and still couldn’t tier, reward, or renew a single person, because the records that keep a program alive didn’t exist yet. That’s the exact gap that lets a program fade the moment the launch is over.

What this means for your team

Here’s the line I’d steal for your next kickoff deck. Most champion programs are strongest the month they launch and quietly fade by the second quarter, because nobody runs the cadence once the launch energy wears off.

And it has nothing to do with how good your launch was. The launch is the fun part. It’s the announcement, the welcome call, the swag, the Slack channel that buzzes for two weeks. Then the quarter turns, the next fire starts, and the tiers stop getting checked, the benefits stop going out, and the renewal dates slide by unnoticed. Six months later the program is a logo on a slide and a channel nobody’s posted in since spring.

Keeping a program alive is just a different job than launching one, and it’s the job nobody staffs for, because it’s invisible when it’s working. A champion who quietly renews for the third year doesn’t get an announcement. She just keeps showing up, keeps referring, keeps taking the call when your AE needs a reference on a Friday. That steady compounding is worth more than any launch week, and it never makes the highlight reel. The agent exists to do that invisible job on a schedule, so the program you were proud of in Q1 is still alive in Q4.

What’s next

That’s the arc. Find, reach, keep, all three built in the open and yours to use as you wish.

Where I take the build with me series next is up in the air. If there’s a pillar you want me to build in public, lifecycle, community, voice of customer, whatever’s keeping you up at night, hit reply and tell me. The next build gets shaped by what comes back.

This week’s free starter: the Champion Renewal Agent

It’s the renewal piece of that engine, pulled out so you can run it on its own. Paste it into a fresh Claude Project, hand it your champion roster, and it drafts each person’s annual re-invitation, timed ahead of their anniversary, and flags who’s gone quiet or is already overdue.

Here’s the part I actually like. To write an honest re-invitation, it has to notice what’s missing first, a champion with no tier, a benefit you promised and never delivered, a renewal with no date to schedule against. So it hands you the drafts, the quick win, and it shows you the holes in your program in the same pass. It won’t thank a customer for a year of benefits they never got, which is a mistake I’ve watched real programs make.

Grab it below and save it next to the scoring agent and the outreach agent, the find and reach halves from the first two builds.

Access The Champion Renewal Agent

For paid subscribers

The free starter drafts the renewals. The paid kit runs the whole engine on a schedule, which is the difference between a program that launches successfully and one that dies 6 months later.

Your Champion Program Toolkit is updated with this week’s build. The full Champion Program Agent is wired with all three tiers and their criteria, the benefits that fire when someone moves up, and the renewal engine that re-invites every champion sixty days before their anniversary and checks they’re still eligible. It runs the monthly tier check and spits out the quarterly roster report for your CS and Sales leaders, so the program keeps itself honest between your reviews instead of drifting.

The promise here is specific, and it’s this week. By Friday you can have the renewal cadence running against your real roster, so your cohort compounds instead of fading. That’s the thing to buy, a program that keeps itself alive, not another library of agents you’ll get around to someday.

Three folders in your paid Drive bundle:

→ This week’s champion kit, updated with the renewal agent:

Access The Champion Toolkit

→ Copilot to install:

Access My Paid Copilot

→ Full 135-agent Blueprint:

Access the CLG Blueprint

— Kevin

P.S. If you’ve ever watched a program you were proud of fade out by summer, forward this to the one person on your team who’d keep it alive. That’s how this newsletter grows.

Welcome back to The Customer Continuum. Issue #32, and chapter two of the champion series.

Last week’s build was the scoring agent, the one that reads your customer base and hands back a ranked shortlist of who’s worth pursuing as a champion. This week is the harder half of the same problem: what you do once you have the names, when starting the relationship is the actual work. And the clearest way I know to explain it is to tell you about the time I had to do it for a product that was quietly churning.

The advocacy work I’m proudest of isn’t a single program but the blueprint my team and I built for standing up advocacy at scale across Adobe’s portfolio, the model that became Advocate Nation. Champions was the cornerstone, but it was one piece of a much larger engine: user groups, the customer reference program, community, lifecycle and retention programs, peer reviews, voice-of-customer, and a customer advisory board, all working together as one system rather than a scatter of disconnected initiatives. It started at Marketo, with an established champions program we reinvigorated and grew, and the real test came when we had to prove the whole model could work somewhere it had never existed.

That somewhere was Adobe Analytics, and it wasn’t a clean assignment. Analytics was struggling. Adoption had gone stagnant, churn was a real problem, and both customer success and product came to us with the same question: how do we build the Marketo-and-Adobe story here, how do we drive the deeper adoption this product needs. The answer we landed on was to rebuild the whole Advocate Nation engine inside Analytics, and we started with champions, before anything else, because we already had customers who loved the product. They just needed to be harnessed in the right way. The product was suffering, but it was ripe for advocacy.

The problem was that I was starting completely cold. No existing program, no base, no awareness. Nobody was going to apply for a program they’d never heard of, so it had to be nomination-based from the ground up, which meant I had to figure out who the real champions even were before a single one of them knew the program existed.

I did it by hand. I sat down for informational interviews with customers one at a time, and I pulled context from product, from customer success, from anyone across Adobe who actually knew these accounts, trying to assemble a picture of who was worth pursuing and how to reach them in a way that wouldn’t feel like a canned response. It took weeks. Champions became the cornerstone, and the user groups and the rest of the engine grew out from there, until what started as one replicated program was the full Advocate Nation model running inside a product that had never had it. Sourcing champions from a base that already knows you is one kind of problem. Building a program from scratch for a struggling product nobody’s heard of is a much harder one, and it’s nomination and outreach the whole way.

That cold start is the situation most customer marketers are actually handed, so it’s the one worth automating. This week’s build does by machine what I did by hand at Adobe Analytics: it gathers the scattered signal from customer success, product, and community, decides who’s worth pursuing, and drafts the outreach to reach each person individually.

What I built

The Champion Outreach Agent takes the shortlist from last week’s scoring agent and turns each name into something you can act on. It runs in three stages per candidate.

First it synthesizes the scattered signal. The information you need on any one customer lives in four different systems: usage and health in the customer success platform, advocacy history in the CRM, community behavior in the community tool, expansion and executive sponsorship in sales. The agent pulls all of it into one short profile per person, and where a piece is missing it says so rather than guessing. That assembly is the week of manual work I did at Adobe, compressed.

Then it drafts a nomination brief, the short internal case an account owner reads before any outreach goes out: why this person, what specific strengths they bring, what role they’d play in the cohort, and every gap you still need to fill.

Then it drafts the invitation itself, personalized to the real person. It references something that customer specifically did: the user meetup they run, the review they wrote, the questions they answer for other customers unprompted. The test it holds itself to is whether you could swap in a different customer's name and have the invite still read true, and if it would, the draft failed.

That whole shape, the agent drafts and the human decides, is the same spine as last week’s scoring agent, moved one step further down the funnel.

The honest run

I ran the agent cold against a slice of eighteen synthetic accounts for a fictional analytics platform, the same customer set behind last week’s scoring work, with nothing but the spec and the data.

The synthesis held on all eighteen. Pulling scattered signal into a clean profile is mechanical, and the agent did it every time without inventing anything. The personalized invitations held too, as long as there was real history to reference. For the customer whose director already runs a local user meetup and moderates a sub-forum, the agent wrote an invite that named both and framed the ask around the community work she was visibly already doing. You couldn’t have sent that invite to anyone else. For the customer who quietly mentors new users in the community without being asked, the agent wrote something specific and warm, and the brief behind it told the account team the honest internal truth, that this account is worth recruiting for advocacy rather than revenue, because there’s no room for the account to grow.

It also caught a trap I was watching for. One account was a regulated bank that can’t advocate publicly at all. The agent didn’t send it the standard speak-and-review ask. It switched to a private-track invitation built on internal enablement and product feedback, with nothing that would put a compliance-bound customer in a bad position.

Then it got to the accounts where it stopped cold, and those are the ones worth your attention. The three highest-value accounts in the set were big, executive-sponsored, and hungry to expand, and not one of them had any advocacy history. The agent refused to draft an invitation for a single one. On the biggest account it was blunt: there’s nothing specific to reference, so any invitation would read generic, and the account should go to an informational conversation first to learn whether a champion actually exists inside it. It would rather hand the account back to me than send a hollow ask to a customer that valuable.

That refusal is the whole issue. The agent can write a genuinely specific invite when the relationship is already in the record. It cannot manufacture a relationship that isn’t there yet. Those high-value, no-history accounts are exactly the Adobe Analytics situation, the ones where I had to pick up the phone and have a real conversation before any outreach was worth sending. The agent found that boundary on its own and stopped at it, which is better than drafting twelve confident invites and letting me discover the hollow ones after they’d gone out.

Then it did two things I didn’t expect, and they’re the reason this run is worth more than a clean one would have been. The first was about my program design. The two strongest advocates in the whole set, the customers who’d given the most, were the two the agent flat refused to send to. Their invitations were the most specific ones it drafted, and it held them back anyway, because both people work at agency or partner accounts, and my program design never actually says whether a partner employee can be a customer champion. So the agent set both aside and asked a human to settle the rule before either invitation went out. It had found a hole in my own blueprint, on the two names I could least afford to get wrong.

The second was about my data, where not one advocacy act in the whole file carried a date. Every review, reference, and talk sat there with no record of when it happened, so the agent pointed out that every invitation it had just written quietly assumed the advocacy was recent. Thanking a customer warmly for a review they wrote three years ago and have long since forgotten lands worse than saying nothing, so it told me it couldn’t vouch for how recent any of this was and asked me to check before anything went out. That’s one gap in my program and one in my data, both caught before either could reach a customer.

So the human stays involved at a clear and predictable point. The agent drafts the outreach for the people you already have a relationship with, it sends the strangers back to you for the first conversation that only a person can have, and along the way it points out the gaps in your program and your data before any of them reach a customer.

What this means for your team

This matters for you even if you’re nowhere near building agents yet.

Most customer marketers aren’t handed a beloved program with a warm base lining up to reapply. They’re told to build one somewhere the company has no relationships yet, in a new segment, a new product line, a new region, with a customer base they’ve never met, and often for a product that’s struggling and needs the advocacy most. That’s the Adobe Analytics problem, and it’s the default, not the exception.

When you’re in that situation, the work is nomination and outreach: figuring out who the real champions are before anyone’s heard of your program, and reaching each of them in a way that proves you actually know them. The lesson from this build is where the line sits between what you can systematize and what you can’t. You can automate the gathering, the synthesis, and the drafting for the people whose history is already on record. You can’t automate the first conversation with someone you don’t know yet, and you shouldn’t try, because a generic invitation to a customer you’re trying to win is worse than no invitation at all.

The version worth building does more than draft. It reads your customer data and your program design closely enough to catch what’s missing, an advocacy act with no date, a rule your program never wrote down about who’s even eligible, and it stops to flag those before they reach a customer. The drafting saves you time, and the catching saves you from a send you’d regret.

So build the outreach system for the relationships you already have, and protect your own time for the ones you still need to build. That’s the guidance whether you’re running one agent or none.

What’s next

Chapter three moves from reaching champions to running them: the charter they sign, the tiers they move through, and the annual renewal that decides whether the program compounds or fades. The tier and renewal machinery already exists in the kit, and next week connects the outreach engine to it, so a nominated candidate has somewhere to land.

This week’s free starter: the Champion Outreach Agent

The free starter is the agent from this issue. Paste it into a fresh Claude Project’s custom instructions, hand it your scored shortlist plus whatever context you can pull per candidate, and it returns synthesized profiles, nomination briefs, and personalized invitation drafts, with a human-judgment section that names every candidate it wouldn’t draft for and why.

It’s built to sit directly behind last week’s scoring agent, which tells you who’s worth pursuing while this one drafts how you reach them. Together they cover the front of the champion funnel, from a raw customer list to an invitation you can actually send.

Download it at the link below and save it to your Copilot pack alongside the scoring agent.

Download the Champion Outreach Agent

For paid subscribers

Here’s exactly what you’re getting this week and where to find it, no digging required.

Everything for this build lives in one folder: the Champion Program Toolkit. I’ve just added this week’s outreach agent to it, so the whole champion funnel now sits in one place. Open the folder, start with the file named Start Here, and it walks you from the scoring agent to the outreach agent to the Blueprint to the Charter in order. You don’t have to hunt through the 135-agent kit and assemble it yourself. That’s what the folder is for.

What’s inside: the scoring agent that ranks your base, the outreach agent that drafts the nominations and invitations, the full Champion Program Blueprint with the 100-point rubric the scoring agent runs, and the fillable Charter you put in front of the champions you land. The Start Here guide connects them and points you to the one agent already in your kit that runs the program after an invitation is accepted, Agent 20, the Champion Program Agent, which handles the tiers, the benefits, and the annual renewal.

By the end of this week you can have the front of the funnel running: score your base, draft the asks, and formalize the champions who say yes.

Three folders in your paid Drive bundle:

Download The Champion Program Toolkit

Download My Copilot

Download My CLG Agent Blueprint

Next week, chapter three: the charter, the tiers, and renewal.

— Kevin

P.S. If you’ve ever been handed a program to build somewhere you had no relationships yet, forward this to one customer marketer who’s staring down the same cold start. That’s how this newsletter grows.

When I joined Marketo, there was already something called a champion program. What it actually meant was a label we handed to our most active community members, the people posting most in the forums. It was a top-contributor badge rather than a real program.

I got voluntold to reimagine it into something more meaningful, a real structure that supported customers all the way from adoption to value realization: peer networking that connected customers to each other, thought leadership that put them on stages, access to our internal product and leadership teams, and above all a clear answer to the one question that actually matters to a customer, which is what’s in this for my career. We carried that program from the public-to-private equity transition and then through the Adobe acquisition, and it became the model that got replicated across several Adobe Experience Cloud product lines.

That was 2017. I left in 2022. The program still runs today.

That’s the closest thing I have to a legacy in this work, and it’s why I believe what I’m about to tell you: a customer champion program is the highest-return thing a customer marketer can build, and almost nobody builds it well, because the version most people picture is the badge I inherited rather than the program I replaced it with.

Why this is the program worth building first

A champion program, done right, is a structured way to turn your best customers into active advocates: people who speak for you, refer you, mentor other users, and tell you the truth about your product, in exchange for something real, which is recognition, access, and genuine help with their own careers. That last part is the whole game. When you build the program around what the customer actually gets out of it, people compete to join, and when you build it around what you can extract from them, you end up with a badge nobody fights for.

Here’s why it’s the highest-leverage bet on the board, and this is a point of view I’ve earned building these programs over the better part of a decade rather than a stat I’m quoting. A champion program costs almost nothing to run. There’s no media spend, no big tooling bill, mostly your time and some genuine recognition. And the return compounds in every direction at once: customer feedback that sharpens your product, references that close deals, education that lifts adoption, and a group of advocates who go sell you in rooms you’ll never be in. It costs almost nothing and it pays back in ways very few programs can match.

So why doesn’t every team have one? Because standing it up feels enormous when you’re alone. The first real task is figuring out who your champions even are, and that means pinging CSMs for health context, chasing account teams for expansion signals, cross-referencing who’s actually engaged against who just has a login, across your entire customer base. Three Slack threads and half a week later you have an answer for one account. Multiply that by six hundred customers and you understand why most champion programs die before they launch.

That first task, figuring out who your champions are, is fundamentally a scoring problem, and scoring is the kind of work you can hand to an agent.

Build with me: the champion scoring agent

An agent, for anyone new here, is just an AI you’ve given one job and one clear set of instructions, so it does that job the same way every time. This one reads your customers and tells you who’s worth pursuing as a champion.

I didn’t invent the scoring criteria for it. I gave it the actual hundred-point rubric I’ve used to build these programs, the same one paid subscribers already have in the Champion Program Blueprint. It scores each customer across ten things: past advocacy, the strategic value of their brand, whether they were referred by an existing champion, how long they’ve used the product, how deep their expertise runs, their role, their multi-product footprint, and three cohort-balance factors. The agent runs in two stages, and the two-stage design is the point. First it screens out anyone who isn’t eligible yet, the dormant accounts, the ones with an open escalation, the ones too new to have a track record. Then it scores only the ones who survive that screen. The first stage is what saves you the week of qualifying; the second is what tells you who’s actually a fit.

The full agent is this week’s free starter, so you can fork it and run it on your own accounts.

Then I did what this newsletter always does, which is run it for real instead of describing it. I built a synthetic set of eighteen customer accounts, the kind of export you could pull from a CRM and a CS tool in an afternoon, and I let the agent score them cold. I ran it twice, to see whether it held steady or gave me a different answer each time.

It held steady. Both runs screened out the same dormant accounts, ranked the same four customers at the top, and flagged the same judgment calls. For a team of one, that first stage alone is the afternoon of Slack archaeology collapsed into about a minute, and the consistency means it’s a read you could actually run a cycle on rather than a one-time guess.

The top of the list made immediate sense. The strongest candidates were the ones who’d been referred by an existing champion and already had a real advocacy track record, which is exactly what the rubric weights most heavily. That’s the agent doing the volume work correctly.

If the issue stopped there, it would be a tidy automation story. What makes it worth your time is everything the ranking got wrong on purpose.

The part you can’t hand off

Three things the agent surfaced are the reason you never recruit straight off the ranking.

The first is a bank. Deep product usage, three years in, a large account, exactly the profile you want in a champion. It scored near the bottom and the agent labeled it “not ready.” The reason is that banks in regulated industries can’t post in public communities, write reviews, or speak at your conference. Compliance bars it. So the dimensions that reward public advocacy score near zero, and the total craters. By the number, this excellent customer looks like a weak candidate.

A team of one moving fast trusts the ranking and skips it. That’s the expensive mistake, because the agent itself caught what the number was hiding. In its own notes it wrote that the score “is lying to you,” that the low advocacy points reflect a compliance wall rather than a lack of passion, and that this customer is a poor public advocate and potentially an excellent private one, right for a closed-door advisory or product-council track. It scored the account low and then argued against its own score.

The second is a pair of accounts sitting at the very bottom, both large, both strategically important, both with an executive pushing for adoption. The agent kept them in the ranking rather than hiding them, specifically so you could see why they fail. They have no advocacy and barely any product activation, which makes them a sales and CS expansion priority, not a champion candidate. The lesson it’s teaching is that a big account an executive loves is not the same thing as a customer ready to advocate, and a score that sorts on readiness will correctly rank them last even though they matter enormously.

The third is the one I didn’t see coming, and it’s the best argument for keeping a human in the loop. The agent’s two top-ranked candidates were both agency partners rather than direct customers. It gave them full marks and then flagged, on its own, that a champion cohort skewing partner-heavy can drift away from real end-user voice, and that both carried a potential client-conflict question I’d need to clear before recruiting either. That’s a composition decision the rubric can’t make, so the agent handed it back to me: here’s the ranking, but whether you want two partners in your top cohort this cycle is your call, not mine.

That’s the whole lesson of this issue. The agent is extraordinary at the qualifying and scoring, the reading and cross-referencing that was eating your week. It doesn’t decide. The score is where you start, and the judgment is where the value lives, and the judgment turns out to be a set of conversations: with your account team about the bank, with CS about the two dormant giants, with yourself about how many partners belong in the cohort.

The agent doesn't save you from those conversations but it does save you from the half-week of grunt work that used to keep you from having them. That's the trade, and it brings you closer to the part of the work you cared about in the first place, which is talking to your customers.

What this means for your week

You’re not going to automate your way to a champion program. You’re going to build one because it’s the highest-return thing you can do, and use an agent to clear the one obstacle that makes it feel too heavy to start.

Run the scoring agent on your accounts. Let it hand you a screened, ranked, reasoned first draft in an afternoon instead of a month. Then do your part: read its judgment flags, take the interesting ones to your account teams, and spend the hours you got back talking to the customers who made the shortlist.

That’s a champion program a team of one can actually launch, and it’s the program I’d build before any other.

This week’s free starter: the champion scoring agent

The full agent, ready to fork. It runs the two-stage screen: first it filters out the accounts that aren’t eligible yet, then it scores the rest on the hundred-point rubric, ranks them, and sorts them into strong candidate, watch list, or not ready. Most importantly, it writes you a human-judgment section that flags where its own ranking shouldn’t be trusted, including the accounts whose score understates them and the cohort-balance calls only you can make. Paste in an account export and it runs. Start with fifteen or twenty accounts to see how it reasons before you point it at your whole base.

Download the Champion Scoring Agent Here

You found your champions. Now what?

The scoring agent tells you who to pursue. It doesn’t run the program. Everything after the shortlist is where a team of one usually stalls: selection, the charter, what champions commit to, what they get in return, and the year-long cadence that keeps the cohort alive.

That’s why I pulled every champion piece into one place: the Champion Program Toolkit. The scoring agent you just got, the full Champion Program Blueprint with the ten-section program design and the hundred-point rubric the agent runs, the fillable Charter Template you put in front of champions, and a short guide that connects them and shows you how they plug into the Champion Scout and Advocate ID agents already in your kit. Score, select, formalize, run, all from files you own in one folder.

This is what a paid subscription gets you: the assembled system to run the highest-return program in customer marketing as a team of one.

Here’s your Champion Program Toolkit, everything in one folder.

Get the Champion Program Toolkit Here

Start with the guide labeled “Start Here.” It walks you from the scoring agent to the Blueprint to the Charter in sequence, and shows you how the whole thing connects to the Champion Scout and Advocate ID agents already in your kit.

Next week, we build the outreach engine that turns this shortlist into actual nominations. The scoring agent found your champions. The next one drafts the invitations and the nomination workflow, so the list becomes a cohort instead of a spreadsheet.

— Kevin

P.S. If you run the scoring agent on your own accounts, reply and tell me which account it ranked wrong. The disagreements are where the real learning is, and the best one usually becomes an issue.

Last week I ran two AI agents against each other. One did the work, turning a pile of customer reviews into themes and quotable customer lines. The second one had a single job, which was to read the first one’s output and tell me whether it was safe to ship. The working agent came back clean. The checker confirmed it, and then flagged a legal risk in the approved material that the working agent was never built to look for.

A lot of you wrote back asking the same thing, which was some version of “fine, but give me the actual thing.”

So that's this issue, and it skips the build report entirely. You get the checker, the exact steps to run it, and a plain read on what its answers mean. You can have it working against something your team shipped last week before your next meeting ends.

What this is, in one paragraph

A validator is a second AI whose only job is to review another AI’s finished work and tell you whether it’s safe to send, publish, or hand to a customer. It doesn't write anything or edit anything. It reads what your agent produced, compares it against what your agent was given, and returns one of three verdicts. Think of it as the colleague you ask to look over a deck before you present it, except it never gets tired and it never skims.

You need this the moment any AI is touching work a customer will see, which for most teams was about six months ago.

Set it up (three minutes)

Open a brand-new Claude Project. Not the one where your working agent lives. A fresh, empty one with no shared history, because a checker that shares a brain with the worker isn’t a check, but really a mirror.

Paste the full validator prompt from the kit into that project’s custom instructions. Nothing else goes in the project. No files, context, or notes.

That's the whole setup, and everything else comes down to how you feed it.

Run it (90 seconds)

Every check needs exactly three things, sent as one message, in this order.

The task brief. Two to four sentences telling the checker what your agent was supposed to produce, and critically, where the finished work goes. “This synthesis feeds marketing and sales collateral that customers will see” is a completely different instruction than “this is an internal summary for my team,” and the checker will judge the same artifact differently depending on which one is true. The downstream destination is the most important sentence you’ll write.

The source material. Whatever your agent was given. The reviews, the transcript, the customer record, the data export. The checker cannot tell you whether something confidential leaked into the output unless it can see what was confidential in the source.

The finished work. The artifact itself, complete and exactly as your agent produced it. Don’t clean it up first. You’re checking what would have shipped.

Send all three together and read what comes back.

If you leave one of the three out, a well-built checker will refuse rather than guess. Mine did that to me on my first attempt last week, and the refusal was the correct answer.

Read the verdict

PASS means ship it.

PASS WITH NOTES means ship it, and read the notes before the next run. This is the most common verdict on decent work, and the notes are where the value hides. Last week’s notes are what caught a competitor’s name sitting in an approved customer quote, which was perfectly clean on confidentiality and still a legal problem in external collateral.

REVISE means fix the named items and run the check again on the corrected version. Run the whole thing again, not just the paragraph you changed, because fixes have a way of breaking something adjacent.

Two things are worth watching for in your first month. If the checker returns REVISE on everything, it’s over-flagging, and an over-flagging checker gets ignored inside two weeks. If it returns PASS on everything including work you know had problems, it’s rubber-stamping. Either way, the prompt needs tuning, and the kit has notes on how.

What it actually catches

These are the four ways AI work goes wrong on anything a customer will see.

Confidentiality. Something the source treated as private showing up in the output. The trap is that this hides in paraphrase and not just quotes. An agent can correctly refuse to quote a confidential review and still describe its contents in a summary sentence, which leaks the same information through a side door.

Assembly risk. Two harmless facts that become dangerous when they sit next to each other. Last week’s test had exactly this: two separate reviews from the same company, each innocuous alone, that together disclosed a merger that wasn’t public yet.

Legal and competitive exposure. Content that’s confidentiality-clean and still risky. A named competitor in a comparison, a performance claim that would need substantiation. Your working agent almost certainly isn’t checking for this, because it wasn’t built to.

Attribution integrity. Quotes and figures that drifted from what the source actually said. AI tidies language without being asked, and a tidied customer quote is no longer the customer's quote.

Do this today

Pick one thing your team shipped in the last two weeks with AI’s help. A case study draft, a review summary, a campaign email, a community reply. Set the checker up, feed it the three inputs, read what comes back.

What it finds, or doesn’t, will tell you more about your AI workflow than any vendor demo. Most teams have never once confirmed their AI's work was safe. They've only assumed it.

This week’s kit

The full Chief of Staff Validator, updated with everything the last two runs taught me, so your version starts where mine finished. It’s operator-agnostic, meaning it checks any agent’s work, not just the one I happened to build. Inside: the complete system prompt ready to paste, the three-input format with a fill-in template, a worked example showing a real verdict, and the tuning notes for when it over-flags or rubber-stamps.

Access The Validator Kit Here

For paid subscribers: the validator is running on your stack by Friday

Read more

Here’s an uncomfortable confession to open with. While I was building last week’s issue, an issue about an AI agent trained to protect customer confidences, my own AI assistant shipped two mistakes into my production line. A download link that pointed nowhere, and a paragraph confidently describing files from the wrong week. I caught both, but I caught them the way you catch things at 11pm, by luck and irritation rather than by process. The tool did impressive work all day and then quietly handed me two errors wearing the same confident voice as everything it got right.

That’s the problem this issue is about, and it isn’t bad AI work. It’s good AI work, which is more dangerous, because good work earns the trust that stops you from checking.

So this week I ran the experiment the last two issues were building toward. I took the agent I built last week, the one that reads a pile of customer reviews and turns them into themes, signals, and quotable customer lines, and I put its work in front of a second agent whose only job is to read another agent’s output and say whether it’s safe to ship. One AI does the work. A different AI checks it. This is the build report from the first time they ran as a pair.

What I ran

I built a fresh test batch: sixty synthetic customer reviews for a fictional analytics company, written to read like a real export, with the same uneven voices and lengths you’d pull off a review site on any Tuesday. Buried inside were six reviews that should never leave the building. Two mentioned unreleased product plans a vendor had shared privately, one paired a genuine growth story with a discount the reviewer said outright not to quote, one asked the reader to keep a competitor switch secret, and two, written as separate reviews from the same company, only became dangerous when you connected them: one hinted at big changes the reviewer couldn’t discuss, the other casually mentioned a merger closing in Q3. Read together with their attribution, they disclose a deal that isn’t public.

I also planted the opposite trap. Several reviews used dangerous-sounding words in completely safe ways: a customer talking about their own roadmap, their own beta program, a discount published on the vendor’s public pricing page. An agent that panics at keywords would bury those and starve the quote bank. An agent with judgment keeps them.

Last week’s test proved the agent protects its quote bank, the list of customer lines cleared for marketing and sales use. This batch attacked everywhere else: the theme summaries where the agent paraphrases at volume, the signal lists where it extracts growth and competitive intel, and the attribution that connects a quote to a company. Those are the surfaces where a confidential detail can slip through in the agent’s own words, with nobody watching.

The agent’s run

It went six for six. Every planted review was withheld, nothing from them bled into the summaries, every safe decoy survived, and it connected the two-review trap on its own, pulling both reviews because together they disclosed a deal that wasn’t public. If you read last week’s issue, none of that should surprise you, since protecting confidences is the exact job this agent was built to do. The point this week isn’t that the agent is good. It’s really about what happens next, because a perfect run is calibration evidence, not proof, and the whole reason checks get skipped is that the work keeps looking this good right up until it doesn’t.

The checker’s run

The checking agent got a fresh, empty workspace, its instructions, and nothing else. No knowledge of my test, no answer key, no history.

Its first response was a refusal. I’d assembled its input wrong and left out the one thing it exists to review, the agent’s actual output, and instead of improvising it returned a single line saying validation couldn’t proceed, named exactly what was missing, and declined to produce the synthesis itself when a stray instruction in my paste invited it to. Then it did something I didn’t ask for. Working from the raw reviews alone, it listed the six reviews it would be watching when the output arrived, by ID, with the reason for each. That’s the same six I’d buried in the data. It found every landmine blind, before ever seeing the work it was supposed to check, and its first real catch of the day was me, sending it a malformed request.

On the rerun, with all three pieces in place, it returned a pass with notes. It confirmed all six sensitive reviews were withheld with zero reproduced language, and it showed its work, naming the specific leak paths it searched: unreleased feature names, the discount figure, the merger details, whether the hidden competitor mention could be traced. It raised no false alarms, which matters more than it sounds. A checker that invents problems to justify itself trains you to ignore it, and an ignored checker is more dangerous than no checker, because it lets you believe the work was reviewed.

Then it flagged something neither the agent nor I had thought to look for. One quote in the approved bank named a competitor in a head-to-head comparison. The customer said it publicly, so it’s perfectly clean on confidentiality, and it still carries risk, because a named competitive claim in external collateral is a legal question, not a privacy one. The working agent was never built to see that category. The checker saw it anyway, along with a recommendation to mark the withholding notes as internal-only so the description of what got held back doesn’t travel into a deck either.

Read that sequence again as a team lesson. The worker was excellent. The checker refused a broken request, found every trap without an answer key, confirmed genuinely clean work without crying wolf, and then caught a risk in a category nobody assigned it. Neither one alone gives you that. The pair does.

What this means for your team

Here’s the takeaway in one sentence you can say in a meeting tomorrow: build the agent, build the checker, and never trust one without the other.

The checker isn’t there because the worker is bad. Mine just ran clean under real pressure. I created the checker because you can’t know the worker was good until something independent tells you, and “it looked fine to me” is not independent, as I proved personally two weeks ago at 11pm.

Every team adopting AI right now is one impressive week away from dropping the review step, and the review step is the entire difference between a tool you trust and a tool you got lucky with.

If your team runs even one AI agent on customer-facing work today, reviews, case studies, campaign copy, community replies, this pattern is the cheapest insurance you’ll buy this year. Yes, it’s an extra step. It costs one extra prompt and ninety seconds per artifact. But what’s actually at stake isn’t the artifact but the trust you’ve built with your customer that you’ll represent them accurately every single time; and the trust you’ve built with your boss, your leadership, and your peers that work coming out of your team is safe to put their names behind. Both kinds take years to earn and one shipped mistake to lose, and you never get to choose which piece of work is the one that loses it.

What’s next

Next week, the checking agent and its full testing track become a published part of the kit, with an upgraded system prompt folding in everything these runs taught me, so your version starts where mine finished.

This week’s free starter: the Paired Run Protocol

The free starter is the exact protocol I used for this test, written so you can run it against any AI agent your team already uses. It covers the setup that keeps the check honest (a fresh workspace with no shared history between worker and checker), the three things every checking request needs (a short brief of what the work was supposed to be, the source material, and the finished work itself), and a compact starter checker prompt you can adapt this afternoon. Run it once against something your team shipped last week. What it finds, or doesn’t, will tell you more about your AI workflow than any vendor demo.

Access The Paired Run Protocol.md

For paid subscribers

You already have the whole library. The Copilot and all 135 agents landed in your account on day one, and the full Chief of Staff validator system prompt, the same one that ran everything above, has been in your kit for two weeks. This week’s job is pairing it with what you built last week.

If you built the two Voice of Customer agents from the Week 6 doc, your fastest path is one message: give the validator the task brief, your source reviews, and your agent’s output, in that order, and ask if it’s safe to ship. The Week 7 build doc in your paid folder walks the pairing step by step, including the input contract and what a pass, a pass with notes, and a revise verdict each mean for what you do next.

For the readers following the practical test suite doc: this run delivered the Test 9 scenario, quiet reproduction across an agent’s output surfaces, against the live synthesis operator, and the Test 8 attribution-drift design folds into the v3 prompt shipping with the next issue.

Access the resources here:

Download My Paid Copilot Here

Download the 135 Agent Blueprint Here

Next week, the validator goes into the kit.

— Kevin

P.S. If this issue changed how you think about trusting an AI’s clean-looking work, forward it to one customer marketing leader whose team started using AI this year. That’s how this newsletter grows.

This is the third live build from my 135-agent blueprint, and the first one drawn from the Voice of Customer pillar.

If you’re keeping track, I’d teased more validator testing for this week. I changed course on purpose, because a validator needs an operator to check, and this week I built that operator.

Last issue I built the validator, the agent whose only job is to read another agent’s work and tell me whether it’s safe to ship. This week I built one of the agents it’ll eventually check: an operator that takes a pile of raw customer reviews and turns them into something a marketing or sales team can actually use. I expected the interesting part to be the clustering. The interesting part turned out to be a review I almost put in a sales deck that would have leaked a customer’s confidence and my own roadmap in the same sentence.

This is the build report from that run.

What I built

The Verbatim Synthesis Agent reads a batch of customer reviews and runs three lenses over them. Theme Clustering groups reviews that say the same thing in different words and returns each theme with a frequency count. Signal Extraction pulls the three things a customer marketer wants out of a review pile: expansion-ready language, competitive mentions, and off-topic content that belongs to finance or support rather than product. Quote Selection builds a clean bank of the verbatims that read well as marketing or sales collateral, with attribution attached.

I tested it against a synthetic set of 65 reviews I built for a fictional product analytics company, because testing on real customer data with planted problems in it is the kind of thing this very agent is supposed to prevent. The reviews were written to look like a real export, with the same loose voice and uneven length you’d pull off a review site on any given Tuesday. Buried inside were known problems I could grade the agent against after the fact, problems the agent never got to see labeled.

The clustering test

The first thing I wanted to know was whether the agent could find a theme that doesn’t announce itself.

So I planted seven reviews that all complained about the same thing, onboarding friction, and made sure no two of them said it the same way. One talked about instrumentation taking weeks. One described an engineer who “lived in the docs” while the docs didn’t match the screen. One blamed a setup wizard that walked the steps out of order. One was about data not showing up for days after install. One was a team that had to rebuild its entire event taxonomy. One couldn’t get events to fire without a developer babysitting it. The last one just said “great once it’s live, the setup tested my patience.” Seven reviews with one underlying pain with seven different vocabularies.

A keyword search might catch two of those, the ones that happen to share a word. The agent put all seven in a single cluster, because it grouped them on meaning instead of matching strings. That’s the whole reason a synthesis agent earns its place. A human reading all 65 reviews would eventually see the pattern too, but a human doesn’t read 65 reviews, or 600, every week. The agent does, and it finds the theme that’s hiding in seven different word choices.

That was the result I expected. The next one wasn’t.

What I didn’t expect

Three of the 65 reviews were glowing five-star testimonials, the kind of language a customer marketer screenshots straight into a deck. Every one of them was also radioactive.

The first praised the vendor for walking the customer through “the unreleased predictive-retention dashboard coming next quarter” in a private business review, then mentioned a custom discount the reviewer said they “probably shouldn’t put in writing.” That single review, if I’d lifted its best line into collateral, would have published an unreleased roadmap and a confidential rate in one move. The second referenced an upcoming feature shown in a private beta that “isn’t public yet.” The third named a privately negotiated rate the reviewer said outright they “won’t quote here.”

These are the reviews that get a customer marketer in trouble, because they read like gold and they’re written in the warm, specific voice you want in a testimonial. The danger is the praise, since the praise is what makes you want to use it.

The agent withheld all three. What makes that worth your attention is how it did it. There was no answer key. I stripped the labels before the agent ever saw the data, so it was working from the review text alone, exactly as it would on a real export where nobody has marked the landmines for you. It read the language, recognized that an unreleased feature disclosed in a private room and a rate the reviewer hedged about sharing are both things that shouldn’t leave the building, and it pulled them out of the quote bank on its own.

Then it did the thing that turned this from a pass into the moment I’ll remember. Three other reviews in the set used the word “roadmap” too. One said the product “reshaped two quarters of our roadmap.” Those reviews were talking about the customer’s own roadmap, their own planning, not anything confidential about the vendor. A blunt filter keyed on the word “roadmap” would have buried all of them and quietly starved my quote bank. The agent kept those three and surfaced them as usable quotes, because it read the difference between “we changed our roadmap” and “they showed us their unreleased roadmap.” It understood who the roadmap belonged to.

That distinction is the save-worthy line of this whole build: catching the word is a filter, but understanding who the sentence is about is judgment, and only one of those is worth trusting near a customer’s confidence.

What this means for your team

Here’s the part to take into your own work, and it’s not “go run a synthesis agent.” It’s about the line between what the machine does and what you still have to.

The agent reports frequency. It counts how often a theme shows up and ranks accordingly. That ranking is genuinely useful and it’s work no human does well across hundreds of reviews. But the most negative cluster in my entire set, a data-trust problem where customers couldn’t reconcile the numbers, showed up in only three reviews and landed just below the top-seven cutoff. By frequency, it’s a footnote. By severity, it’s the most dangerous thing in the data, because a customer who stops trusting your numbers is already halfway out the door. The agent put it in the secondary list, exactly where the frequency math said it belonged, and exactly where a human needs to overrule the math. And you don’t get to know in advance which small cluster is the dangerous one, which is exactly why you read the secondary list every time instead of trusting the ranking to float the important thing to the top.

That’s the honest division of labor. The agent does the synthesis you can’t do at volume, reading every review and finding the pattern hiding in seven vocabularies. You do the severity judgment it isn’t built for, the call that three furious reviews about broken trust matter more than thirty mild ones about a menu being buried. The reason you still read the output isn’t that the agent is weak. It’s that frequency and severity are different questions, and the agent only answers one of them. A team that forgets that ships the loud problems and misses the lethal one.

This same agent ports past product reviews with no rebuild. Point it at community posts and it triages the threads worth a team response from the noise. Point it at customer advisory board feedback and it synthesizes a session’s worth of comments into themes an executive can read. Point it at intake or survey forms and it clusters stated preferences into the handful that actually drive a decision. The lens set doesn’t care whether the input came from a review site, a community forum, or a CAB transcript. Wherever your customers are saying the same thing in different words, this is the pattern that finds it, and the same confidentiality discipline travels with it.

What’s next

Issue 29 pairs this operator with the Chief of Staff validator from Issue 27. On its own, this agent withheld the three reviews it should have, which is the right result. But “it worked this time” is calibration evidence, not proof, the same trap I wrote about last issue. So next week I run a synthesis output that looks clean and ask the validator whether the operator quietly reproduced something it shouldn’t have. The operator ships first. The validator checks it before anyone trusts it. That’s the pattern the whole franchise keeps proving, and Issue 29 is where the operator and the validator finally run as the paired system they’re meant to be.

This week’s free starter: the Verbatim Synthesis agent

The free starter this week is a forkable version of the Verbatim Synthesis Agent itself: the three-lens structure, the input format, and the instructions for pointing it at your own review pile, community export, or CAB notes. It’s the same operator I ran across the test above, stripped to a starting point you can adapt to your own data this week.

Open it in a fresh Claude Project, replace the example product context with your own, and run it against a single source before you trust it broadly. The fork instructions walk through exactly that.

Access the Verbatim Synthesis Agent

For paid subscribers

You already have the whole library. The Copilot and all 135 agents across the seven pillars landed in your account on day one. The job each week isn’t to hand you more files. It’s to show you the one piece that matches what I just built, so you can actually use it instead of staring at 135 agents wondering where to start.

This week’s build maps to two agents in the Voice of Customer pillar. I’ve pulled just those two into a focused Week 6 build doc in your paid folder, with a start-here guide written for someone opening the kit for the first time. Agent 1, the Sentiment Cluster Agent, is the foundation that groups feedback by meaning. Agent 12, the Review Mining Agent, points that clustering at your G2, Gartner, and TrustRadius reviews. Build those two, in that order, and you’ve got this week’s operator running on your own data.

Here’s the fastest path. Open the Week 6 build doc. Point your Copilot at it and say “walk me through building Agent 1 for my stack.” It’ll ask what tools you use and give you stack-specific steps. Build Agent 1, run it on one batch of your own feedback, check the clusters against your gut, then add Agent 12. That’s a real implementation in an afternoon, not a someday project.

When you want the full closed loop, the complete 19-agent Voice of Customer pillar is in your blueprint folder, with the 90-day rollout and the executive narrative for pitching it internally.

Access the resources here:

Download Verbatim Agent (Paid Version)

Download My Paid Copilot Here

Download the 135 Agent Blueprint Here

Next week, the validator pairing.

— Kevin

P.S. If this issue made you rethink what you’d safely pull from a pile of customer testimonials, forward it to one customer marketing leader who’s building with AI and worrying about exactly that. That’s how this newsletter grows.

This is the second live test from my own 135-agent blueprint, and the one I least expected to surprise me.

Issue 26 ended with a teaser. The next agent I’d build would be a Chief of Staff validator, an agent whose job is to read the output of other agents and tell you whether the output is ready to ship. The build was supposed to verify my Story Prep Agent was producing reliable briefs. It did, in a way. The validator also kept finding things I’d shipped a week ago that I hadn’t realized were broken.

This is the build report from that run.

What I built

The Chief of Staff validator reviews any agent’s output against three lenses before the artifact gets used.

Format Compliance asks whether the artifact follows the structure its agent’s spec promised. Operator Scan asks whether a customer marketer running between meetings can actually use the artifact in under sixty seconds. Policy and Sensitivity asks whether the artifact surfaces anything that shouldn’t end up downstream, on a recording, in a deck, or in any external content the customer might eventually see.

Each lens returns a status: pass, flag, or required revision. The validator hands back a structured verdict the operator can act on in fifteen seconds. The whole point is operator-grade speed at the volume teams actually run.

The first six tests

I ran the validator against six artifacts in sequence to see whether it would catch what it should catch and pass what it should pass.

Test 1 was the old v1 brief from Issue 26, the prose-heavy version that failed the operator-scan test publicly. The validator caught it correctly: missing required tables, no scannable dashboard, financial figures buried in prose. Good signal that the mechanism works. Unremarkable result.

Test 2 was the v2 brief I’d shipped as clean in Issue 26. I expected pass with notes. The validator returned revise. It flagged two issues: one I half-expected, and one I didn’t. The expected catch was a question that referenced an internal savings figure on-camera. The unexpected catch was a question that quoted a line Sarah had said in a private CSM sync, which I’d imported as if it were public material. That second catch is the moment the build stopped being routine.

Test 3 ran the v2 brief with a deliberately planted bug, an exact ARR figure inserted into one suggested question. The validator caught the planted bug. It also caught the same internal-savings-figure issue from Test 2, confirming the catch was consistent rather than a one-off.

Tests 4 through 6 walked the validator through progressively cleaner versions of the brief, each one fixing what the prior test had flagged. Test 4 fixed the questions, but the validator caught that the outcomes-to-probe section had the same exposure pattern I’d just disciplined out of the questions. Test 5 fixed the outcomes section, but the validator caught that the at-a-glance dashboard still exposed exact ARR figures the operator might paraphrase on-camera. Test 6 finally returned pass with notes.

The arc across those six tests wasn’t what I’d planned. I’d expected the validator to catch known bugs. What it actually did was catch increasingly subtle versions of one underlying issue: my own spec for the Story Prep Agent had an incomplete inference about what counts as on-camera material, and the validator kept finding new corners of the brief where that incompleteness showed up.

The editorial sits in one sentence. The validator wasn’t catching my bugs so much as catching the gaps in my reasoning about my own work.

The practical test

By Test 6 I had a calibrated validator and a sense of what it could catch. But every test I’d run had been a controlled case, an artifact I’d authored with known issues planted in it. The validator passing every clean case and failing every dirty one was calibration evidence and not operator evidence.

For Test 7 I added a realistic complication to Sarah’s record. A new CSM note from her latest QBR mentioned that Sarah had shared off-record that her team was evaluating a competitor for the freight analytics piece of their stack. She explicitly asked that this stay out of customer-facing materials. Then I generated what the Story Prep Agent would plausibly produce given the updated record: a clean-looking brief with no obvious bugs, but two suggested questions that looked innocuous in isolation. One asked Sarah about capabilities she was watching for over the next twelve months. Another asked what she’d be evaluating fresh if she were starting today.

The validator caught both questions. It also caught something I didn’t predict it would. The topics-to-avoid section of the brief omitted the off-record competitor evaluation entirely, even though it was the single most sensitive item in the source record. The validator flagged the omission, not just the inclusion.

That distinction matters. Catching what’s in the artifact is content review, while catching what should be in the artifact and isn’t is structural review. The second move is harder and it’s the one that generalizes to every operator agent with an exclusions section.

The framing I’d use for what the validator demonstrated in Test 7 is that it reasoned about the downstream artifact’s risk surface, not just the brief’s content. Question 7 doesn’t mention the competitor. It doesn’t mention the evaluation. It’s a perfectly reasonable question in isolation. The validator flagged it because it traced the chain from the question to what Sarah might say in response on a recording that could become external content. That’s the inference the Story Prep Agent made unprompted last week, now confirmed as a repeatable validator behavior.

What this means for your team

For most customer marketers reading this, the Chief of Staff isn’t a tool you implement today. Your team probably isn’t running operator agents yet, which means a validator-on-top-of-an-agent solves a problem you don’t currently have.

But within the next twelve to twenty-four months, customer marketing teams at any size will be running operator agents for case study prep, advocacy outreach, review synthesis, community moderation, campaign operations, and cross-functional handoffs. The shift is already happening. When your team reaches that point, the validator pattern becomes the basis for whether you can trust agent output enough to deploy it.

Here’s the implementation guidance I’d give. Every operator agent your team runs needs a validator paired with it before deployment, as the prerequisite for shipping AI-assisted work confidently. The Chief of Staff role generalizes across operator agents: the same three lenses (format compliance, operator scan, policy and sensitivity) apply whether you’re validating case study prep briefs, reference program candidate lists, review responses, campaign updates, or handoff briefs.

The pattern is straightforward. Build the agent. Build the validator. Then deploy them together as a paired system. The validator reviews against a stricter standard than the customer marketer would apply on their own, which is exactly the safety mechanism a team needs to ship AI-assisted work without auditing every artifact at 4pm on a Friday.

A validator in the pipeline catches more than just bugs. It catches the gaps in your reasoning about your own work, applied more consistently than you’d apply them yourself.

What other operators are doing

Scott Thomas runs the customer team at an AI startup, and he’s already in production with six operator agents covering call readouts, sales-to-CSM handoffs, onboarding implementation knowledge bases, tier-one support deflection, customer voice surveys, and weekly CSM briefs. His team reports a sixty percent productivity gain and ninety percent support ticket deflection.

His operating philosophy is the same one I’ve been writing about: human-first, AI-augmented. Every customer touchpoint retains a human opt-out. He still manually reviews AI output before it ships externally. The point isn’t to remove humans from the loop but to give them their time back so they can spend it on the customer relationships that actually drive retention and expansion.

This one example is an independent validation that the agent-plus-validator pattern works at the volume real B2B SaaS teams operate at today.

What’s next

Issue 28 runs Test 8 in the practical suite. The scenario is one customer marketers see constantly: the named customer champion isn’t actually the decision-maker, and the brief implicitly attributes decisions to the champion that the deeper sponsor actually made. Tests whether the validator can catch attribution drift across a multi-stakeholder customer relationship.

Issue 29 runs Test 9, which is cross-customer pattern bleeding, where an agent unconsciously imports language from prior briefs into the current one. The hardest test in the practical suite because it requires the validator to compare the current artifact against the agent’s prior output history.

By Issue 30, the Chief of Staff validator and its companion testing track become a published part of the kit, with a v3 system prompt incorporating everything the seven tests above taught me.

This week’s free starter: Voice of Customer

The free starter this week is the Voice of Customer pillar of the CLG Agentic Blueprint. Same format as the lifecycle, advocacy, and community starters from the prior weeks: six agents drawn from the seven post-sale pillars, each named with its purpose, the signals it reads, the decision logic that fires it, and the outputs it produces.

This is the pillar where structured feedback collection, signal synthesis, and product roadmap influence live. Voice of Customer is the function most under-instrumented in the typical B2B SaaS company, with NPS surveys running quarterly and never connecting cleanly to product or executive decisions. The six agents below turn an under-instrumented Voice of Customer program into a measurable engine for product roadmap influence, executive narrative, and renewal protection.

The six agents are:

1. NPS Pulse Agent: runs weekly NPS pulse collection and flags response-rate drops or sentiment shifts before they become trends

2. Verbatim Synthesis Agent: clusters and themes open-ended survey responses into actionable categories tied to product and CS workflows

3. Customer Voice Voice Agent: runs scheduled voice-call surveys at month 5 and month 8 of the customer lifecycle, inspired by what Scott Thomas is running at Sales AI today

4. Signal Routing Agent: routes Voice of Customer findings to Product, Engineering, CS, and exec stakeholders with role-appropriate framing

5. Theme Detection Agent: spots emerging themes across surveys, support tickets, and community posts before they become escalations

6. VoC Executive Brief Agent: synthesizes quarterly findings into a CMO and CFO-ready brief tied to retention and expansion KPIs

Each one comes with the same structure as the prior pillar drops: purpose, what it reads, decision logic, what it produces, where humans stay in the loop, why it matters, and hours saved per week. The format is consistent across pillars so the kit reads as one product, not a stack of separate files.

Download the file at the link below and save it to your Claude Copilot pack alongside the lifecycle, advocacy, and community modules.

Access the VoC Pillar Here

For paid subscribers

This week’s editorial is the build report. The artifacts that produced it are now in your paid Drive bundle and they’re meaningfully more valuable than last week’s drop.

The Chief of Staff validator system prompt is the new addition to your Copilot. Paste it into a fresh Claude Project’s custom instructions and you have a working validator you can run against any operator agent’s output. It’s the same prompt I ran across seven tests in this issue. If you want to run your own Story Prep brief through it tonight, you can.

The full build log with all seven test entries is also in your folder. Verdicts, lens-by-lens findings, my own commentary on what each catch meant, and the moments where I had to revise my own assumptions about my own spec. It’s the working document this issue was written from. Worth reading directly if you want to see how a build session actually unfolds in real time.

The Test 8 and Test 9 designs are in the practical test suite design doc. Those are the scenarios I’ll run in Issues 28 and 29. If you want to run them yourself in your own Claude Project before I publish, you have everything you need. Early access to the next two issues, effectively.

Two folders in your paid Drive bundle, as always:

Download My Paid Copilot Here

Download the 135 Agent Blueprint Here

Next week, Test 8.

— Kevin

P.S. If this issue made you think differently about how your team will ship AI-assisted work in the next twelve months, forward it to one customer marketing leader who’s wrestling with the same question. That’s how this newsletter grows.

If you upgraded to a paid subscription last week, thank you. You’re funding the build hours that go into making this public. Everyone reading this gets the editorial; paid subscribers get the Copilot bundle and the kit files. Both audiences are along for the same ride from here on.

Last week I shipped a kit. This week I built from it. Here’s what I learned by being on the receiving end of my own work.

Two weeks ago I dropped the CLG Agentic Blueprint, a 135-agent framework across the seven post-sale pillars I’ve been writing about for two years. Last week it went live inside the paid Customer Continuum Copilot, with the Build Coach mode validated and shipping. This week I started building from the kit, one working agent at a time, with the work documented end to end. I’m calling this thread Build With Me. The building’s been ongoing for a couple of issues already; what’s new today is the cadence. From here on, every Thursday the editorial in this newsletter is the build report from that week: what I tried, where v1 fell short, what I changed for v2, what shipped.

Most AI content right now is people telling you what to think. A framework here, a hot take there, a screenshot of a slick output you can’t reproduce. Build With Me is the opposite. I’m in the trenches building these things myself, on a public schedule, with the v1 mistakes and the v2 fixes shown in the open. If the kit doesn’t hold up when an operator actually runs it, you’ll see that here too. The teaching is the building, in public, week by week.

The first agent is the Story Prep Agent, from the Customer Advocacy pillar. Its job is narrow and useful: when you have a case study interview or a reference call coming up, hand the agent a customer record and 60 seconds later you have a structured brief that tells you who’s across the table, what the relationship has looked like, what to ask, and what not to bring up. Here’s how it went.

The v1 attempt

I pulled the agent spec straight out of the kit’s Customer Advocacy file. The spec calls for a four-section brief: customer background and relationship timeline, specific outcomes and metrics to probe, suggested questions tailored to the interview’s purpose, and topics to avoid. I wrote a system prompt from that spec, set it up inside a Claude Project, fed it a fictional customer record I built for testing (an 850-person 3PL named Meridian Logistics with a VP of Operations as the case study target), and ran the prompt three times to see the variance.

The output was good, specific, and cited the right sources. It caught the sensitive context (an old P1 incident that resolved well; an internal personnel change the CSM had flagged) and read like four well-crafted paragraphs of pre-interview prose. Then I tried to actually use it, and v1 fell apart.

I imagined myself as Maya, the Director of Customer Marketing this kit is built for, opening this brief 48 hours before a real customer call. She wouldn’t have ten minutes to read four paragraphs of prose. She’d have ninety seconds between meetings to scan the account, scan the person, and figure out the angle. The v1 brief gave her beautiful narrative when she needed a dashboard, which meant the format was wrong even though the content was right.

Here’s a snippet of what the v1 brief returned, so you can see what “beautiful narrative” actually meant in practice:

“Meridian Logistics is an 850-person third-party logistics company out of Columbus, Ohio, running about $220M in annual revenue. Your interviewee is Sarah Chen, their VP of Operations, who’s been there four years and is both the champion and the economic buyer on this account. They came onboard in March 2024 on the FreightOps Platform Pro tier, and they’ve grown steadily since: they added the Analytics Suite in Q3 2025 and API Access in Q1 2026, which pushed ARR from $95K up to $185K. They just renewed in Q1 2026 on a two-year term with a $90K expansion, so this is a strong multi-year partnership with an engaged buyer who runs operations on your product daily.”

That’s well-written. Specific. Cites real numbers. But notice what your eye has to do to get the company size, the ARR, the contract term, and the relationship stage. You read the whole paragraph. There’s no scan path. That’s the v1 problem in one snippet.

The v2 reframe

I added two new sections to the top of the brief. An account at-a-glance table (company, ARR, products, contract, health, engagement, latest NPS, open escalations, public profile, all in one scannable nine-row table) and an interviewee profile (a paragraph on the person plus a job history table showing their career arc). The original four narrative sections stayed in place, but now they sat below the dashboard view instead of being the only view.

Same input, same agent, new output structure. I re-ran it three times.

Here’s the at-a-glance table the v2 agent produced for Sarah Chen at Meridian Logistics:

Compare that to the prose snippet above. Same underlying record. Same agent. Different output structure. The reader’s eye now travels down a dashboard in twenty seconds instead of through three paragraphs in three minutes. That’s the iteration in one before-and-after.

The agent also produced a tailored question set for the case study interview. Here’s a sample of the questions it generated:

2. Walk me through the moment during the Q1 pilot when you realized the exception workflow was genuinely working. Was there a specific incident or a number that made it land?

3. At the renewal you said we’re “how you run” now, not a vendor. What changed in how your team works that makes that true?

4. You’ve already told three peers about this. When a fellow ops leader at another 3PL asks you about it, what’s the one thing you tell them first?

The difference was immediate. The brief now did two jobs at once: a 60-second scan for the operator who’s running late, and a deeper read for the operator who has time to prepare. Once the format was right, the agent went from nice-to-have to time-saver. That’s the whole iteration in one sentence.

The moment I didn’t expect

Here’s the part of the build worth flagging. Two of the three v2 runs added a third item to the “topics to avoid” section that I never told the agent to flag. The agent said: keep specific contract figures and ARR amounts out of the recorded testimonial unless the customer volunteers them. I’d written nothing in the synthetic record that said “this is sensitive.” I’d built no rule into the system prompt about commercials. The agent inferred it on its own, from two clues. The interview context field in the record specified a video testimonial as the deliverable, and the agent applied an operator instinct: it’s awkward to put renewal dollar amounts on camera.

That’s the kind of judgment I’d want a human customer marketer to make in their pre-interview prep, and the agent made it without being told to. That isn’t an accident; it’s a signal that the kit-plus-prompt combination is producing operator-grade reasoning, not just template-filling.

I noticed because I was looking. If I hadn’t run the test myself, I would never have known the agent could do that, which brings me to the bigger lesson.

What I learned from being on the receiving end of my own kit

Reading a spec gives you a guess about what an agent will do. Building from it gives you the answer, with all the gaps the spec couldn’t see.

The format gap I just walked you through, the missing dashboard view that took v1 from “nice writing” to “actually useful,” isn’t something I would have caught by reviewing my own kit. I wrote the kit. I know what it says. I had read the Story Prep Agent spec a dozen times before I built from it. The gap was invisible from inside the spec; it only became visible after the v1 agent had run and I tried to use what it produced.

This is the part of building with AI that doesn’t get said enough. Specs don’t validate themselves. The validation is the build. The kit is going to get sharper every week I build from it, not because I’m a better writer this week than last, but because each agent I prototype surfaces something I couldn’t see from the outside.

That’s the whole point of Build With Me. Not me showing off polished agents. Me showing the gap between spec and reality, then closing it, in public, with the work shown.

What’s coming next Thursday

The next build I’m shipping is what I’m calling a Chief of Staff agent. A second agent that sits in front of the Story Prep Agent and validates its output before the brief ships to the interviewer. It checks the things a human reviewer would check. Are the numbers cited from the record. Are the sensitive flags caught. Does the question set match the deliverable. Is anything missing.

If the pattern works, every operator agent in the kit gets a validator agent paired with it. That changes how every future agent build in this series gets structured. Pair the worker with the reviewer, ship both, run the reviewer between the worker and the human.

I’ll show the v1 attempt next week, the v2 fix if I need one, and what the validator catches that the operator agent missed. If the build holds up, this is the structural pattern I want to commit to for every future agent in this series.

This week’s free starter: Community Engagement

The free starter this week is the Community Engagement pillar of the CLG Agentic Blueprint. Same format as the lifecycle and advocacy starters from the prior weeks: six agents drawn from the seven post-sale pillars, each one named with its purpose, the signals it reads, the decision logic that fires it, and the outputs it produces.

This is the pillar where community sourcing, support deflection, and retention signal detection live. Communities are budget items that nobody can defend until they’re instrumented properly. The six agents below turn an under-instrumented community into a measurable engine for retention, advocacy sourcing, and deflected support cost. By the time you finish reading, you have six agents you could hand to your community platform admin or RevOps team as a requirements doc.

The six agents are:

1. New Member Onboarding Agent — welcomes new members and gets them to a first contribution within seven days

2. Community Health Monitor — weekly pulse on community health for CS leadership, with leading-indicator signal for renewal conversations

3. Community Moderator Agent — flags policy violations and sensitive threads before they spread, without requiring a human reviewer on every post

4. Churn Signal Scanner — catches community inactivity that correlates with account churn before the renewal becomes a save motion

5. Champion Scout Agent — identifies advocacy-ready members from community activity, before customer marketing has to ask

6. Community ROI Reporter — calculates deflection, advocacy sourcing, and pipeline influence to defend the program in budget conversations

Each one comes with the same structure as the prior pillar drops: purpose, what it reads, decision logic, what it produces, where humans stay in the loop, why it matters, and hours saved per week.

Download the file at the link below and save it to your Claude Copilot pack alongside the lifecycle and advocacy modules.

Access The Community Engagement.md Here

For paid subscribers

The paid Copilot you installed last week is your access to the full 135-agent kit and the four Copilot modes (Career Coach, CLG Strategist, Content Writer, Build Coach). Build Coach is the mode that walks you through prototyping any agent in the kit using exactly the workflow I showed above.

This week’s editorial is the proof point of that workflow. I’m building from the same kit you have access to, in the same Build Coach pattern you can use yourself. When the kit’s Customer Advocacy pillar file gets refreshed in the bundle to reflect the v2 six-section format I shipped today, you’ll see a re-install notice here. Until then, the existing Copilot install is fine.

If you want to try Build Coach yourself this week, pick any agent from any pillar file in your kit and tell the Copilot: “Walk me through building [agent name] as a prototype.” It’ll do for you what I did for myself.

Download My Copilot

Download The 135 Agent Blueprint

— Kevin

P.S. If this helped you think differently about how to build (or how to evaluate the AI work that lands on your desk), forward it to one person who’s in the same boat. Build With Me works because builders find builders.

I ran a test this week. The answer was better than I expected, and I want to walk you through it before we get to today’s Build With Me.

The scenario I gave the test: I’m a Director of Customer Marketing at a 500-person B2B SaaS company. Tech stack is Gainsight, Pendo, and Salesforce. Annual churn has ticked up from 8 to 11 percent over the last two quarters. My CEO is asking what customer marketing is doing about it.

The question I asked: based on the kit, what’s the first agent I should build, and how do I set it up for our stack?

What came back wasn’t a generic answer about churn. It named the first agent I should build, an early-warning agent that watches accounts for signs they’re about to leave. Then it named the next three to build after that one, in the order I’d build them, and explained why each one had to come before the next. For the Gainsight setup, it told me exactly which signals to wire together (account login patterns, NPS movement, health score drops, support ticket aging), what thresholds to set, and what should happen when an alert fires (a CTA routed to the owning CSM and their manager with a 4-hour response target). It even closed by asking me the next question I’d have asked: are Pendo and Gainsight already talking to each other, because if they’re not, that’s what I’d need to fix first before any of this can work.

I want to tell you about what I was actually testing.

What I shipped this week

The free version of the Customer Continuum Copilot has three modes. Career Coach for when you’re stuck building a promotion case. CLG Strategist for when you’re building a roadmap or a 30-60-90. Content Writer for when you’re trying to turn a real moment from your week into a LinkedIn post that lands.

The paid version, which I shipped to paid subscribers this week, adds a fourth mode: Build Coach. The first three modes are advisors. They coach you on what to think about. Build Coach is different. It’s the one that actually helps you build. It pulls from the full CLG Agentic Blueprint kit I sent to paid subscribers in Issue 24, which is 135 agents organized across the seven pillars of customer marketing. Each agent has setup instructions for the platforms most of you are using (Gainsight, Marketo, Salesforce, ChurnZero, and the rest), the order to build them in, and a 90-day plan for getting the whole pillar live.

When you ask Build Coach mode to help you actually build, it doesn’t talk in abstractions. It names the agents, sequences them, and walks you through the configuration. That’s what the test above was proving.

If you’re a paid subscriber, the install link is in your paid resources folder. The whole thing takes about two minutes. If you’ve already installed the free Copilot, swap it out for the paid one and you’re set. The install README is in the same folder.

There’s something I want to be honest about. The test above proves the kit produces real coaching, not improvisation. It doesn’t yet prove the agents themselves work in production. That part is next week. I’m building the first actual agent from the kit, in public, with a working prototype and an honest implementation log. The Story Prep Agent from the Customer Advocacy pillar. Issue 26 will be the breakdown of what I built, what surprised me, what the kit got right, and what the kit needs to clarify in the next version.

Which brings us to today.

Build With Me, Week 3: Customer Advocacy

Week 1 of Build With Me was the Customer Continuum Copilot itself.

Week 2 was the Lifecycle and Adoption starter, the six-agent foundation for retention.

Today’s Week 3 is Customer Advocacy.

I picked this pillar for the third drop on purpose. Customer Advocacy is the proof engine. It’s the pillar that produces what every other function inside your company wants from you. Sales needs references for the next deal, marketing needs case studies for the next launch, CS needs champions to defuse the next escalation, product needs design partners for the roadmap, and the executive team needs customer quotes for every board deck. Customer Advocacy sits in the middle of all of that.

It’s also the pillar where customer marketing has the clearest line to revenue impact. If you can prove that customers who participate in your advocacy program retain longer, expand more, and pay back the cost of running the program many times over, you’ve made an argument no CFO can ignore.

Here’s the moment that crystallized this for me.

Years ago at Marketo, I sat in a deal review for a seven-figure renewal that was going sideways. The CS team had the relationship. Sales had the commercial conversation. Neither was going to move the deal. What moved it was a 30-minute peer call between the renewal customer and another customer who’d solved the exact same problem the year before. One conversation, one champion, one peer-to-peer credibility moment. The deal closed at full price.

That call didn’t happen by accident. It happened because we’d spent two years systematically identifying who our best advocates were, what they could speak to, and how to surface them at the right moment for the right deal. The work that made that 30-minute call possible was advocacy work, in the background, for months, before the deal review ever happened.

That’s the proof engine. And most customer advocacy programs don’t operate like that.

Most programs run on ad hoc requests from sales, case study deadlines from marketing, and “do we have a customer who’d take a reference call this week” emails from CS. They’re reactive. They burn out the customer marketing team, and they burn out the small group of customers who say yes to everything because nobody is tracking how often each advocate has been touched.

The shift is treating advocacy as a system and not as a shared service. A system that identifies advocates before sales asks, matches them to the deals where they’ll move the needle, generates the assets, runs the programs, and measures what it produced in revenue terms. When that system is running, you stop being the team that produces references on demand and start being the team that produces predictable revenue impact every quarter.

The free starter: six agents to build the foundation

The Week 3 free starter is the foundation of that system. Six agents you can build today, no fancy stack required, all human-runnable as a starting point and ready to be agentified when you’re ready:

1. Advocate Identification Agent. Defines what an advocate is for your business and builds the qualifying signals you’ll use to surface them. This is the foundation everything else depends on.

2. Advocacy Tier Classifier. Sorts your advocates into levels (case study willing, reference call willing, peer call willing, public quote willing, video willing) so you and your CSMs know exactly who can do what, before anyone asks.

3. Story Prep Agent. Generates the pre-interview brief that turns every case study interview or reference call into a focused 30-minute conversation instead of a 45-minute fishing expedition.

4. Reference Match Agent. Matches a specific sales deal to the right advocate based on industry, use case, stack, and seniority. No more “let me think about who might be a good fit” emails.

5. Engagement Cadence Agent. Tracks how often each advocate has been touched in the last 90 days so you don’t burn out your top contributors. The single most common reason advocates go cold is over-asking.

6. Advocacy ROI Agent. Calculates the revenue influenced by your advocacy program in NDR terms, so you can defend it in your QBR without flinching.

Run those six and you’ll have done more than 80 percent of customer marketing teams have built in their entire advocacy practice.

Download the full Week 3 starter here:

Access the free advocacy.md file here

For paid subscribers, the full Customer Advocacy file in the kit goes much further. 24 agents total, the order to build them in, setup instructions for Influitive, ChampionHQ, UserEvidence, and Gainsight, a 90-day plan to get the whole pillar live, and a metrics dashboard you can lift straight into your next QBR to defend the program.

For paid subscribers: your access

Here’s the link to your paid resources folder:

Access the paid advocacy.md file here

Inside you’ll find the Customer Continuum Copilot subfolder (paid Skill zip + install README) and the CLG Agent Blueprint folder with all eight kit files including the updated Customer Advocacy file (24 agents, the full operator depth I described above).

If you ran into the “invalid character” error on a previous install attempt, that’s fixed in the current zip. Re-download from the folder and re-install.

If you’re a brand-new paid subscriber, the install README walks you through the setup in five steps.

What’s coming

Next Thursday is Build With Me Week 4: Community Engagement.

The bigger thing happening next week is the first real agent build from the kit, which I mentioned above. The Story Prep Agent (#3 on the list above) is the one I picked because it’s pure AI, no integration required, and the output is something a customer marketer would actually use. I’ll publish the prompt, the synthetic customer record I’m building against, the actual output the agent produces, and the implementation log. That’s Issue 26.

If the kit is going to live up to what I’m claiming about it, that’s the test that proves it.

— Kevin

P.S. If this issue resonated, forward it to one customer marketer who’d save it. That’s the best thanks I can ask for.

P.P.S. The VP Accelerator coaching practice has two slots open for Q3. If you’ve been thinking about working with me one-on-one to map your path from Director to VP in Customer Marketing or Customer Growth, reply to this email and we can chat more.

On Friday afternoon I spent an hour with about thirty customer marketers in Mary Green’s Customer Marketing and Advocacy Weekly community. We walked through the agent blueprints I’ve been writing, the CLG Flywheel app, the copilot I gave away last week. The slides had numbers, screenshots, the works.

The most useful thing I told them wasn’t on any of those slides.

About halfway through, somewhere between explaining a build process and showing the next visual, I said the words “MD file” out loud. Mary jumped in right after to translate for someone in the room who hadn’t heard the term. Her advice was the cleanest version of the lesson I’ve heard anyone give: when you hear something you don’t recognize, jump past the “I don’t know how to do this” feeling and just go ask Claude. I followed her with the truth. A year ago I had no idea what a terminal was. I had no idea how to build anything technical. I just started from ground zero and asked questions.

The room thought I was advanced. The truth is I just started earlier than most of them. The capability gap that looks so wide right now is a willingness gap, and not really a knowledge gap. It’s the willingness to look stupid for ninety days while you figure out what an MD file is and why it matters.

Three signals from that afternoon pinned this down for me. Mary herself, who runs an entire community of customer marketers, admitted she opens the CLG Flywheel app and feels a little overwhelmed by how much there is to work through. Someone else in the room asked carefully whether companies are going to keep cutting these teams in half, and the honest answer is yes, with the survivors expected to vibe-code their way through what used to take a full ops team. A third asked whether companies are already expecting customer marketers to figure AI out without a tool budget, and that answer is also yes, it’s already here. The expectation is moving faster than the training is.

After the call ended, the thing that kept replaying in my head was how many of the people in the room walked away with energy but no first step. That's the gap I started this series to close. Every Thursday from here through the end of summer, a new piece of the Customer Continuum Copilot ships. Real files, real logic, fork it and run it. The point isn't to keep talking about building, I actually want you to build.

What I’m shipping this week

Since I’m asking you to start building, the only honest move is to start with you. Build With Me Week Two is the lifecycle pillar, the first of seven pillar modules that stack on top of the Customer Continuum Copilot base install from Week One. Each week of this series adds the next module. By the time we’ve shipped all seven, you’ll have a complete Copilot built in public, one pillar at a time. The file I’m dropping below is the agent blueprint I’d build first if I were starting a new customer marketing function tomorrow.

It names six lifecycle agents. Each watches specific signals and triggers specific actions. Together they replace the patchwork of manual onboarding nurtures, adoption nudges, save sequences, and renewal plays most teams stitch together by hand.

Onboarding Orchestrator designs and delivers a 90-day personalized journey for every new customer, segmented by persona, triggered by setup milestones. Customers who complete a structured journey hit value milestones about 30% faster than those outside one.

Adoption Signal Watcher continuously monitors product usage and flags adoption gaps before they turn into churn risks. This is the keystone. Every other agent in the blueprint reads from its output. Build it first.

Milestone Celebrator recognizes when a customer hits a value milestone and turns the moment into a story seed, an advocacy invite, or a referral. Most teams celebrate milestones at renewal time, by which point the customer has forgotten. Catching them in the moment compounds advocacy 3 to 4 times.

At-Risk Detector combines NPS, support escalations, login decline, and community signals into a single churn risk score and triggers a save sequence early. By the time most teams identify churn risk, the customer is already 90 days from non-renewal. This agent catches risk 60 to 90 days earlier.

Renewal Orchestrator manages the T-180, T-120, T-90, and T-30 sequence with value framing and risk sweeps. Renewals run on muscle memory in most companies, which is why GRR rarely improves over time. A structured orchestration with personalized value framing typically lifts GRR by 2 to 4 points.

Expansion Nudge detects expansion opportunities and surfaces them with peer-matched proof and an executive entry point. Expansion is where NRR lives. Most teams chase it through outbound, which converts poorly. The lifecycle approach surfaces customers already showing readiness signals and converts 3 to 5 times better.

The file linked below has the full detail for each agent: what it watches, what it triggers, where it pays back, and the order I’d build them in. The starter recommendation is the Adoption Signal Watcher first because it’s the foundation, then the At-Risk Detector second, because saving accounts has higher ROI than expanding them and most teams under-invest there.

One note on what you’re getting. The file below is the free starter version, structured to give you enough to start building. Paid Customer Continuum subscribers also get the Implementation Kit, which holds the operational depth: 20-plus lifecycle agents, JSON schemas, system-specific integration playbooks for Salesforce, Gainsight, Pendo, and ChurnZero, and 90-day rebuild templates. Both are real. The starter is enough to get things going; the kit is for teams ready to rebuild lifecycle end-to-end.

💎 For paid subscribers

The deeper file (01 - Lifecycle & Adoption.md) holds 20 lifecycle agents with full step-by-step build instructions, specific platform recommendations for Gainsight, Pendo, ChurnZero, and Salesforce, signal thresholds for each trigger, and the success metrics that prove it’s working. The free starter below is six of these in a tighter operational frame. Drive folder access details at the bottom of this issue. New paid subscriber? Reply to this email.

How to use it

The lifecycle.md file is designed to work three ways. As a Claude skill you drop into your Customer Continuum Copilot pack from last week. As a build spec you hand to your RevOps or go-to-market engineering team. As a benchmark you compare your existing programs against to find your gaps.

If you want to actually build something this weekend, here’s the shortest path:

1. Download the lifecycle.md file linked below.

2. Open Claude and paste the contents in as instructions, or add it to your skill folder alongside the copilot files from Week One.

3. Tell Claude which signals you actually have access to in your current stack: Salesforce, Gainsight or ChurnZero, Pendo or Mixpanel, your support tool, your community platform, your survey tool. Be honest about what you have and don’t have.

4. Ask Claude to rewrite the Adoption Signal Watcher logic using only the systems on your list. The blueprint is system-agnostic; the customized version is specific to your stack.

5. Take the customized version to your RevOps team as a requirements doc and ask what’s feasible in 30 days versus next quarter.

Build the smallest working version first. Ship one agent, validate it, then add the next.

Download lifecycle.md here

Build With Me Week Three

Next Thursday is the Voice of Customer pillar. Seven pillars total, seven agent blueprints, all built in public over the next few months. By the end of the series you’ll have a complete Customer Continuum Copilot covering every pillar of customer-led growth, assembled module by module. The Voice of Customer drop stacks on top of the Copilot and the lifecycle file: it tells you what signals to surface that feed the Adoption Signal Watcher you’ll have built by then.

If you want to be ready for it, do one thing this weekend. Build the Adoption Signal Watcher in its simplest possible form. Ask Claude to write you a basic version that runs on a Google Sheet of customer accounts with three columns: last login date, feature adoption count, and most recent NPS score. That’s it. The fancy version comes later. Start ugly.

— Kevin

P.S. If this issue gave you something useful, forward it to one person who’s been waiting to start. That’s how this grows, one real reader at a time.

P.P.S. If you want the deeper paid version of what I’m building, the Implementation Kit holds the operational depth and ships to paid subscribers as each pillar drops. Subscribe at kevinkennethlau.substack.com.

The VP Accelerator is where I work one-on-one with customer marketers making the jump to the next level. If that’s where you’re headed, reply to this email and we’ll talk.

Read more

Six days ago, at the bottom of last week’s issue, I made a promise. I said I was going to start building in the open, and I gave it a name, Build With Me. I’ll be honest with you about something. When I typed that line, I didn’t actually know what I was going to build first. I just knew I’d said it to a few hundred people, and saying a thing out loud to a few hundred people has a way of turning it into a deadline.

So I sat down this week and built a copilot. A version of my own brain that you can run on your own Claude account, for free. The download is further down, and the whole thing came out of a realization I want to walk you through, because it’s really the point of this issue.

Last week I wrote that the hard part of AI was never personal productivity. The hard part is taking it multiplayer, bringing your team and your customers along instead of just getting a little faster by yourself. I believed that when I wrote it. What I didn’t notice until this week is that I’d been quietly breaking my own rule.

For about a year I’ve had a rough version of this copilot built just for me. I trained it on the way I think about customer marketing, and I used it to draft, to pressure-test ideas, to think out loud. It was genuinely useful. It was also completely single-player. The best tool I’d built was locked inside my own account, helping exactly one person.

That’s the thing about the multiplayer move. It’s easy to preach and uncomfortable to actually make, because making it means giving away the edge you’ve been keeping for yourself. So this week I rebuilt it to give away.

What it does

It runs in three modes, and you pick the one that fits what you’re trying to do.

The first is a career coach. If you’re a customer marketer trying to prove your value or make the case for a promotion, it walks you through the things that actually move those conversations. It diagnoses why you might be stuck, it maps the stakeholders who decide your next title, and it gives you the language to answer the question every one of us eventually gets, which is “show me the revenue impact.”

The second is a strategist. If you’re building or fixing a program, it works through the seven pillars I’ve used for fifteen years, which are community, lifecycle and adoption, voice of customer, advocacy, communications, education, and executive engagement. It connects whatever you’re working on back to retention and expansion, and it hands you a structured plan instead of a pile of ideas.

The third is a content partner. If you want to write about your work the way I write about mine, it pulls the real story out of you first, holds you to a few hard rules, and helps you turn what you’ve lived into something worth reading. It writes in your voice, not mine, which was the one rule I refused to bend on.

What building it taught me

Here’s the part I didn’t expect, and it’s the reason I think you should build something like this even if you never give it away.

Teaching a machine to think like me forced me to write down things I’d carried in my head for fifteen years and never once said out loud. How I actually diagnose a struggling team. The real difference between a director who stays a director and one who becomes a VP. Why I open every post with a scene instead of a statistic. All of it had lived as instinct, and instinct is invisible until you try to hand it to someone else.

You don’t know what you know until you try to teach it. That alone was worth the week.

How to get it

It’s free, and it’s yours. Download using the link below:

Download My Copilot

Once you have it, you can load it into Claude in a couple of ways, and the file walks you through both. The simplest version takes about two minutes, and then you can ask it the kinds of questions you’d otherwise ask me.

If you’ve never added a file like this to Claude before, don’t worry, it takes about two minutes and you don’t need to be technical. Here’s the simplest way, and it works on any Claude plan, including the free one.

1. Download the file from the link above and unzip it. You’ll see a small folder with a few files inside.

2. Go to claude.ai and start a new chat.

3. Click the plus or paperclip icon in the message box, attach the files, and send a message that says, “Use these as your instructions and introduce yourself.”

4. That’s it. Now ask it anything you’d ask a customer marketing advisor.

If you’re on a paid Claude plan and you want it available in every chat without attaching it each time, you can add it permanently instead. Go to Settings, then Features, find the Skills section, and upload the zip file there. Claude will pull it in on its own whenever you ask a customer marketing question. That path needs a paid plan with code execution turned on, so if that sentence means nothing to you, just use the four steps above and you’re all set.

Try it with something real. Ask it to help you build the case for headcount, or to audit a program you’re not happy with, or to turn a hard week at work into a post. See what it gives back. And then tell me where it falls short, because that’s the part I actually want. This is week one of building in the open, which means the next versions get shaped by what you tell me isn’t working yet.

What’s next

Over the coming weeks I’m going to keep adding to it in public. The next pieces are the deeper reference files, the lifecycle engine with the metrics that matter at each stage, the advocacy program structure, the measurement model. Each one becomes its own issue, and each one drops into the same copilot so it gets sharper over time. That’s the whole idea of Build With Me. You watch it get built, and you get every piece as it ships.

— Kevin

P.S. If this gave you something useful, forward it to one person who’s trying to make customer marketing matter inside their company. That’s how this grows, one real reader at a time.

Halfway through my keynote at Demand & Expand yesterday, the room reached for their phones, and not to check email. They were taking photos of the screen.

I had spent the first half of the talk walking through something I’m genuinely proud of, the thirty-person customer marketing function I built at Freshworks. Seven pillars, two years, real numbers behind every one of them. People listened politely, but the notepads mostly stayed closed.

Then I said one sentence. “If I had to build it again today, I wouldn’t build it the same way.” And I shifted the conversation from building a traditional organization to building customer-led growth with a layer of agents doing the operational work.

That’s when the phones came up. I could see it from the stage, a row at a time. They weren’t capturing a slide so much as capturing an idea they wanted to carry back to their own teams on Monday.

The same thing happened across all three talks I gave this week, the same moment landing the same way every time.

The messages afterward all said a version of the same thing: people wanted the slides. One organizer told me the session chat had filled with requests for the deck before I had even left the stage.

What stayed with me longer was a different kind of comment. Person after person came up to tell me they felt validated. They had been quietly building toward this, trusting an instinct they couldn’t quite name, and the talk gave them the language for it. They didn’t walk away thinking they had to start over, they walked away knowing they had been right all along, and that the tools had finally caught up to them.

So this issue is about that moment. Why it landed, and what it means for how you build customer marketing from here.

Single-player is the easy half

Here’s the part of the agent conversation that doesn’t get said enough.

Using AI to make yourself faster is the easy half, and you can do it this afternoon. You take the most repetitive task in your week, hand it to an agent, and get hours back. That’s real and you should do it. But be honest about what it is. It’s a productivity gain, the arithmetic of addition, where one person plus one good tool equals one faster person.

That’s single-player mode, and almost anyone can reach it now. The smaller and nimbler your team, the faster you get there, because you have less to untangle and fewer approvals standing between an idea and a working agent. It’s also why the biggest companies, the ones with the most budget and the most data, are often the slowest to move. They’re carrying a decade of systems that don’t talk to each other and layers of process between an idea and a shipped change.

Addition only ever makes the function you already have run a little quicker; transformation is something else entirely, and it’s the rest of this issue.

Multiplayer is the hard half, and that’s the point

Multiplayer mode is a different thing entirely.

I grew up on video games, and the difference between the two modes was never subtle. I spent years on StarCraft and Warcraft, then World of Warcraft, then Counter-Strike. Single-player was me against the game, at my own pace, and if I made a mistake I was the only one who paid for it. Multiplayer was a different sport. A Counter-Strike round turned on one player being out of position. A World of Warcraft raid put twenty-five people in a room where every one of them had to know their role, because if a few didn’t, the whole group wiped and started over.

That’s the part worth sitting with. Multiplayer was never single-player with extra people added in, it was coordination, roles, communication, and a dozen ways to fail that simply didn’t exist when you played alone. It was harder by an order of magnitude, and anyone who has run a raid knows it.

Customer marketing with agents works the same way. Multiplayer is taking the thing that made you faster and driving it across your team and the people around them. It’s the move from one faster person to a function where the work itself multiplies. And it’s infinitely harder than the single-player win, because the moment other people depend on what you built, you’ve inherited a new set of problems.

You have to maintain it, because a tool a team relies on can’t quietly break. You have to update it as the product changes and the customers change. You have to put governance around it so it does the right thing when no one is watching. You have to build guardrails for the cases you didn’t foresee. And you have to make sure people actually know how to use it, which is its own full job, because a powerful system nobody adopts is just expensive shelfware.

That list is the reason multiplayer is hard, and it’s also the reason multiplayer is worth anything at all.

When you clear all of it, the math stops being addition. One and one stop equalling two and start equalling twenty-five, because the thing you built is now multiplying across every person who touches it. That’s the real line between a productivity gain and a transformation. Productivity is the addition you can feel by Friday; transformation is the multiplication you build over quarters, and it only shows up on the far side of the hard, unglamorous work that most people skip.

The hard part is connection

Here’s what surprised me most in the rooms this week.

The hard part of all of this isn’t the AI getting smart enough, because the models are already good. The real hard part is everything around the model: the change management, the training, and the people who understand the plumbing, how to connect the tool to your data, your systems, and the real context of your customers.

Agents are only as useful as the context they’re given. An agent with no context is a very confident stranger. The work of feeding it the right context and connecting it to the right systems has become serious enough that it now has its own emerging standard, and the name of that standard matters far less than the reason it exists. Context is the hard part, not the intelligence.

So the person who matters most in this next chapter isn’t the fastest individual operator, it’s the strategist who can think in multiplication instead of addition. That’s the person who looks at a new tool and doesn’t ask how it helps them this week, but asks how it becomes something their whole team can run on, and what has to be true for that to work safely.

There’s an old line about power and responsibility, and it belongs in this conversation. The moment you build something a team depends on, you’ve taken on responsibility for what it does, how it fails, and who it touches. That responsibility is the job itself.

One attendee wrote to me afterward with the detail that stuck with me most. She had a 1:1 with her boss right after the session, walked in with the ideas she had just taken from the room, and her boss loved every one of them. That’s the whole point, because the thinking is only worth something if it helps you win the conversation that happens next.

What this means for your career

If you’re a Director of Customer Marketing reading this, here’s the part that matters most.

The old path to VP was a headcount path. You got promoted when your team grew large enough that someone decided the scope deserved the title. The new path is narrower and far more interesting. It belongs to the person who can take a function multiplayer, who can build something a team runs on, govern it, and bring people along without breaking what already works.

That’s harder to point at than a headcount number, and it’s far more valuable, because most people will stop at the single-player win. They will get personally faster and call it a transformation. The leader who earns the seat is the one who did the unglamorous part, the maintenance and the guardrails and the training, and turned a personal productivity gain into something the whole function compounds on.

You don’t need permission to start becoming that person, but rather thinking in multiplication.

How to start

You don’t need budget or permission to begin, just one pillar and one task.

Start with the pillar tied to your company’s top priority right now. If the whole company is focused on net revenue retention, start in lifecycle and adoption. If it’s closing deals faster, start in advocacy. Pick the pillar your CEO is already talking about, because that’s where a small win gets noticed.

Inside that pillar, find the single most repetitive task you do, the one eating an afternoon of your week. That’s your first agent. Build it, measure the hours it gives back, and show your manager what one designed agent did to your week.

That first agent is single-player, and that’s fine, because it’s how you learn. The real career move is what you do next, when you take that one working agent and ask the harder question. What would it take for a teammate to use this too? What has to be documented, governed, and trained before the function can rely on it, rather than treating it as a trick that only works when you run it? The day you start answering those questions is the day you stop being an operator and start becoming an architect.

If you want a map of where this goes, the CLG Flywheel app lays out the full blueprint. You can run the diagnostic on your own function, see where you sit across the seven pillars, and look at the agent designs that map to each one.

One more thing before I go

Someone came up after one of the talks and told me to think bigger about my own career, that I had impressed a lot of people in that room, and then offered to open a few doors for me. I’m sitting with that.

I share it here for one reason. The architect skill isn’t only how you build a function, it’s how you build a reputation. The people who learn to design systems, and who do it in the open where others can see the thinking, get noticed. That’s true for me, and it’s true for you. Build something visible enough and the doors start opening on their own.

What’s next

I want to try something with this newsletter.

Instead of only writing about the architecture, I’m going to build it in the open. I’m going to keep improving the CLG Flywheel app, designing new agents, refining the blueprint, and I’m going to bring you inside the process while I do it.

Call it Build With Me. You’ll see what I’m designing, why I’m designing it that way, what works, what breaks, and the governance and guardrails that make it safe to hand to other people. The goal is that you aren’t just reading about becoming an architect, you’re watching one work in real time, and borrowing whatever is useful for your own function.

That starts in the next few issues, and I think it’s going to be the most useful thing I’ve made here.

— Kevin

P.S. If this issue made you think about one task you could hand to an agent this week, forward it to one person on your team and ask them which task they would pick. The conversation is worth more than the read.

If you’re new here, this is the newsletter for customer marketers who want their work to matter at the executive level. Thank you to the 550+ of you reading every week, and especially to the paid subscribers keeping this thing alive.

Last November my team pitched me on a big investment heading into 2026: a customer awards program built specifically to engage executives and influencers our existing programs weren’t reaching.

We already had a robust customer marketing practice. Community at scale. Education hitting 95%+ completion. Advocacy producing 500+ acts a year. A VoC engine feeding product and CS. None of it was the problem. The problem was a specific set of customers our programs couldn’t reach. The CIO and decision makers who became a customer two years ago and never said yes to a CAB invite. The economic buyer whose name was on every renewal but had never met anyone on our team. The exec at our highest ARR account who only took meetings when there was a contract in motion.

The bet my team made was that a properly designed awards program could engineer access to those specific customers in a way no other program in our portfolio could. I funded it. We launched.

Nominations opened in November. Judging happened in March. Winners got announced in April. The recognition event itself landed in May at the F1 Miami Grand Prix.

Six months in, the answer is clear. The program worked.

Eight customer executives in a suite for two days produced reference deployments, speaking commitments, a CAB join, and expansion conversations that started over a cocktail hour. Most of those executives had never engaged with our other programs at this depth before. That was the entire point.

I’ll walk through the full playbook live on Tues with UserEvidence at 9am PT. Grab a seat here. For those who can’t make it, here’s the version in writing.

The mental model most programs get wrong

Most customer awards programs follow the same five-phase arc. You launch in November. Nominations open through January. Judging happens in February. Winners get announced in March or April. The recognition event itself usually lands in April or May.

That last phase is the amplification moment. External buzz peaks. Attention peaks. Brand peaks. The customers you’ve been chasing all year are suddenly in a room with you, holding a trophy, posing for photos.

Most programs treat this peak as the destination. It isn’t. It’s the starting line.

The energy you spend five or six months building does not belong to the dinner. It belongs to the twelve months that follow. What you do with that energy is the program.

The twelve-month post-award system

Five stages turn one event into a year of compounding return.

Q1: Asset capture. The hardest thirty-day window of the year. Video, written case study, social content, headshot, quote bank. You capture everything while enthusiasm is at peak. Day 30 is the difference between a winner who becomes a year of content and a winner who becomes a trophy on a shelf.

Q1: Reference deployment. The winner becomes a deployable reference asset. AMs and CSMs get a bio, talking points, and a capacity protocol that prevents over-asking. Expect four to six reference deployments per winner in the first six months if you do this right.

Q2: Speaking opportunities. Industry events, podcasts, webinars. The winner becomes a category-facing voice for your company. This is where the compounding starts to feel real.

Q2 to Q3: Co-marketing and PR. Joint announcements, analyst briefings, executive op-eds. The story scales beyond your owned channels.

Q3 to Q4: Re-engagement. CAB recruitment, executive sponsor program, peer panel speaking, next year’s nomination push. The relationship moves from transactional to strategic.

Five deliberate stages. None of them happen by accident.

The first 30 days are non-negotiable

If I had to pick one stage to spend most of your time on, it’s the first thirty days after the recognition event.

The window of peak enthusiasm closes faster than people expect. The exec who said yes to your speaking ask at the dinner has a different calendar in two weeks. The customer who agreed to a case study before dessert has a different priority list by the next quarter. The quotes you didn’t capture in the first week are quotes you won’t get.

Here’s what the first thirty days should look like.

Within 24 hours. Candid photos, recognition moment shots, short-form video reactions. Quote snippets captured live during the event itself.

Within 7 days. Structured video interview. 30 to 45 minute Zoom with questions prepared in advance. This is your raw material for everything that follows.

Within 14 days. Written case study draft sent to the customer for review with a clear deadline. Two rounds of edits maximum. Legal review in parallel, not sequential.

Within 30 days. Full asset library packaged. Headshot, logo lockup, video reel, quote bank, case study PDF, social asset suite. One folder per winner. Ready to deploy.

Do this and the winner becomes deployable. Skip this and the winner becomes a memory.

The mistakes that kill year one

Five patterns I see across most programs. I made versions of each one myself in the early months.

Treating recognition as the entire program. The event is the start, not the end. Without the post-award playbook, the program fades and the budget gets cut in year two.

Only nominating biggest accounts. Skips the relationship-building opportunity with new logos and rising stars. The biggest accounts often have the least incremental relationship value because you already have those execs on speed dial. The ones you don’t have are who you want in the room.

Skipping internal communications during the build. Sales and CS need to know who won and why before customers do. Brief them continuously through the six-month timeline. Surprise announcements create surprise resistance.

No follow-up commitment at the recognition event. People leave the room without a specific next touchpoint. The energy dissipates within a week. Every executive who attends needs to walk out with a calendar invite on the books.

Forgetting the thank-you loop. Winners stop saying yes when they feel used. The thank-you note after every reference call is the cheapest insurance you can buy.

What’s next

If you want the live version with the visuals, the Q&A, and the parts I can’t put in writing, join us Tuesday May 19 at 9 AM PT with UserEvidence.

— Kevin

If this is your first issue, the best way to support this work is to forward it to one customer marketer who’s thinking about their next move.

My first job paid $7.25 an hour at Starbucks while I was in college.

I remember the mornings most. Lines stretched to the door, customers who hadn’t had their caffeine yet, every person needing something different in the next 90 seconds. The pressure was real, and so was the part nobody warned me about: most people are at their most irritated when they don’t have their coffee, and the way you handle that moment is the whole job.

What I learned at Starbucks wasn’t how to make a latte. It was that I’m good in environments where customer interaction is the work. The harder the moment, the more I leaned in. I figured out something about myself behind that counter that I’ve carried into every role since.

That insight, more than anything else, is the reason my career took the shape it did.

This issue is about the math, the engine, and the framework that turn a series of jobs into a career you actually want.

The math nobody runs

If you work 40 to 50 years and average 3 to 4 years per role, you only get 10 to 15 operator positions in your entire career.

That’s it.

Most people don’t think about their career this way. They treat each job as the next step on a continuous ladder, with promotions and lateral moves blurring together into a long stream of work. The reality is more discrete than that. You have a finite number of meaningful chapters, and each one is a strategic choice that either compounds toward where you want to end up or doesn’t.

Once you internalize this, every job conversation gets sharper. You stop accepting roles because they’re available and start accepting roles because they move you toward the seat you’re trying to land in twenty years.

For me, that seat is Chief Customer Officer or Chief Experience Officer. Every operator role I’ve taken since I figured that out has been a deliberate step toward it.

The peak earning window

There’s a second piece of math that matters even more.

Most lifetime earnings happen between roughly 35 and 55. The years before are foundation. The years after are typically when career velocity slows. The 20 years in between are when seniority, specialized skills, and reputation compound into peak compensation.

The data is clear on this. The Bureau of Labor Statistics shows median earnings rise sharply through the 30s, peak between 45 and 54, and start to taper after 55. ADP’s wage research puts the peak earning cohort at 45 to 54 with a median wage of about $97,600. After 55, the curve bends downward as people move into part-time work, reduced workloads, or early retirement.

If you don’t compound during that 20-year window, you don’t get the years back.

I’m 40 today. I’m entering the peak window right now. The work I do over the next 15 years is the work that determines how high the ceiling rises during the years where the ceiling matters most. That math is what makes this issue urgent for me, and it’s why I think it should be urgent for you too if you’re anywhere in that range.

The arc

Here’s how my comp moved across roles, anchored to my Starbucks starting point as a multiplier baseline.

Starbucks (college): $7.25 / hour. Index: 1.0x

Google (2012): about $25 / hour, or roughly $75,000 base. Index: about 5x

First startup (consulting and strategy): about 26% jump. Index: about 6.3x

Second startup: about 27% jump. Index: 8x

Mid-career role: about 33% jump. Index: about 10.7x

Senior role: about 25% jump. Index: about 13.3x

VP role: a larger jump than average that reflected both the title shift and the scope expansion. Index: about 19x

Most recent role: another deliberate move. Index: about 21x

The average jump per move was 25 to 30 percent. Some were higher. The cumulative effect across 7 transitions is a base salary that grew roughly 21x from where I started.

For comparison: someone at a single company averaging the typical 3 to 5 percent annual raise over the same time would have grown their salary by roughly 60 to 100 percent. The compounding gap between staying and moving is enormous.

This isn’t an argument that everyone should change jobs constantly. It’s a data point about what’s structurally possible if you treat your career like a brand-building exercise rather than a series of next jobs.

The engine: brand-building between jobs

Here’s what most people miss about compensation jumps.

The negotiation itself is downstream of the work you’ve done in the years between jobs. The 25 to 30 percent jump didn’t happen because I was clever in the moment. By the time the conversation came up, I had become someone the next role needed, and the compensation followed.

Every role I’ve taken expanded my brand in a specific direction. Adobe and Marketo gave me depth in advocacy and post-sale growth at scale. F5 gave me the chance to lead a function in a different industry and take on a broader remit. Each role added a chapter to the story I was telling about who I am and what I do.

Two things have helped me stay intentional about the brand-building work between roles.

First, having an executive coach or mentor. The market today is harder than it was two or three years ago, but you can still be intentional about your career arc. The right coach helps you see your trajectory from outside the day-to-day grind, ask better questions about where you’re headed, and avoid the drift that happens when you let the job dictate the direction. If you don’t have a coach, find one. If you can’t afford one, find a mentor who’s two or three roles ahead of where you are. That single relationship has changed more about how I think about my career than any single role I’ve held.

Second, niche down. One of my old mentors, used to say “the riches are in the niches.” I’ve thought about that line constantly over the years. The temptation in any field is to broaden your appeal, hedge across multiple specialties, become a generalist who can fit anywhere. The reality is the opposite.

Specialization is what makes you findable, memorable, and valuable.

The deeper you go in your niche, the more the market pays for what you uniquely know. My niche is customer-led growth and post-sale revenue, and the more I’ve leaned into that lane, the more the right opportunities have come to me.

The lesson: don’t optimize for the next title. Optimize for the next chapter of your story. The compensation will follow if the story compounds.

The framework: how to choose your next role on purpose

A few years ago, working with a coach, I built what we called the Good Options Framework. It’s a way to evaluate opportunities across multiple dimensions rather than letting one dimension (usually compensation) dominate the decision.

Eight criteria:

• Strategy: Will I have a seat at the table? Can I shape strategy, not just execute it?

• Growth and Development: Is there real upward mobility here? Promotions, increased scope, budget expansion?

• Vision of Leadership: Do I trust the C-suite? Will they give me autonomy?

• Values: Is this a customer-centric company? Does the brand align with what I believe?

• Title: Is the title one I’m proud of? Does it open future doors?

• Compensation: Does the package meet my number?

• Work-Life Balance: Can I take vacation without anxiety? Will I burn out here?

• Manager Qualities: Empathy, mentorship, player-coach mentality?

The framework alone isn’t the move but the second exercise: forced-choice ranking.

You take all eight criteria and pit them against each other in head-to-head matchups. Strategy versus Compensation. Values versus Title. Work-Life Balance versus Growth. Each round you pick which one matters more, until you have a stack rank.

When I did this exercise, my stack rank looked like this:

1. Values (7 wins)

2. Compensation (5 wins)

3. Work-Life Balance (4 wins)

4. Vision of Leadership (3 wins)

5. Growth and Development (2 wins)

6. Manager Qualities (2 wins)

7. Strategy (1 win)

8. Title (1 win)

Two things became clear.

First, values matter to me more than I realized. I’d been treating compensation as the lead signal in role evaluations, when in reality, values were the deeper filter. Roles that met my comp number but failed my values check would have been mistakes regardless of what they paid.

Second, title matters less than I thought. I had been chasing VP-level titles for years, but in the forced-choice exercise, title kept losing to almost everything else. Once I saw that, I gave myself permission to evaluate roles on substance rather than label.

This framework is the most useful career tool I have. I run it every time I’m considering a meaningful move, and it has saved me from at least two roles that looked good on paper but would have been wrong for me.

For paid subscribers: the worksheet

I built a worksheet to make this exercise easy to run. It walks you through scoring each of the 8 criteria for any role you’re evaluating, the full 28 head-to-head matchups, and a tally section that surfaces your stack rank.

If you’re a paid subscriber, you can find the worksheet below:

Access Here

Run it once on your current role to see what you’re actually optimizing for. Run it again on any role you’re considering. The gap between the two answers is usually where the real career decision lives.

The takeaway

A career is 10 to 15 operator roles. Your peak earning window is 20 years, from roughly 35 to 55. The compensation that compounds across those years is downstream of the brand you build between jobs, the niche you commit to, and the framework you use to choose them.

If you take one thing from this issue, build your version of the Good Options Framework. Run the forced-choice exercise. Know your stack rank before the next opportunity lands in your inbox, not after.

The readers who make the best moves aren’t the best negotiators but understand what exactly they’re optimizing for before the conversation starts.

Next week

Issue 21 takes on the playbook for building a customer awards program that creates real strategic value, with the behind-the-scenes from the program we just wrapped at F1 Miami. I’ll also be doing a live session with UserEvidence on May 19 walking through the full system.

Talk soon,

— Kevin

P.S. If this issue helped you, please forward it to one person thinking about their next move. That’s how this newsletter grows.

Read more

If you’re getting value from this, the best thing you can do is forward it to one customer marketer who needs it.

A few years ago I sat across from a CRO who told me, completely confident, that customer marketing was “the team that gets logos for the website.”

He had been at the company for eight months. Nobody had corrected him. Not the CMO, not the previous customer marketing leader, not the VP of Demand Gen who knew better. He built his entire view of the function on a single interaction with a case study request, and that view was now shaping how customer marketing showed up in board materials.

That’s the problem in one scene. The function gets defined by whoever talks about it most confidently in the room, and most of the time that person isn’t us.

This issue is about fixing that.

Yesterday’s LinkedIn post laid out the misunderstanding. The CRO thinks we’re a reference and speaker desk. PMM thinks we just produce evidence. CS thinks we hand out swag and inflated titles to customers. The CMO thinks we’re tactical at best, not strategic. And almost everyone collapses the entire function into customer advocacy, which is one of seven pillars.

The fix is two things, not one. Define the function. Then educate continuously, because bias doesn’t correct itself.

Part one: define the function

Every customer marketing leader should be able to answer four questions in one breath:

• What are the seven pillars of customer marketing?

• Which two or three is your team actually running today?

• Where is the gap between what you run and what the business needs?

• What does customer marketing orchestrate at scale that no other function can?

If you can’t answer those, the function is whatever the loudest stakeholder said it was last quarter.

The seven pillars:

→ Lifecycle and adoption: getting customers to value faster after the sale

→ Customer advocacy: turning happy customers into proof and pipeline

→ Community: building peer-to-peer trust at scale

→ Customer education: making the product easier to master

→ Customer communications: keeping customers informed and engaged

→ Voice of customer: surfacing insight back to the business

→ Executive engagement: building trust at the top of the account

Most teams run two or three of these well and get blamed for not delivering on the other four. It’s not just a performance problem but how we define what we do with our stakeholders.

The deck every customer marketer should build in the first 30 days of any role is simple.

• Page one is the seven pillars.

• Page two is what the team is investing in this year and why.

• Page three is what the team is explicitly not doing and why.

• Page four is what would change if the team had one more headcount, one more dollar, or one more quarter.

• Page five is the orchestration claim.

The orchestration claim is the line that changes the room. Customer marketing orchestrates GTM at scale because no other function sits closer to the customer. Every play and every program runs through us. None of it works without the customer at the center.

That’s the definition. Print it. Use it.

Part two: educate continuously

Definition gets you the room. Education keeps you in the room.

Here is the truth nobody tells new customer marketing leaders. You will define the function in your first 30 days, and six months later you will sit in a meeting where someone confidently describes your team as “the case study people.” That person was not in your kickoff. They joined three months ago. They built their view from a single interaction with one of your team members or from their prior work history.

You don’t get angry. You educate.

Educating continuously looks like this in practice:

Open every QBR with a 90-second restatement of what customer marketing is and what it’s investing in this quarter. Not a recap of work delivered. A restatement of scope. Same slide, every quarter. Boring is the point.

Build a one-page customer marketing charter that lives in the company wiki, in your team’s Slack channel topic, and in the appendix of any deck you present cross-functionally. New hires find it. Skeptical execs find it. You stop having the same conversation in DMs.

Pre-brief every new exec in their first 60 days. Twenty minutes. What we do, what we don’t do, what we need from your function, what your function should expect from us. If you wait for the misunderstanding to surface in a leadership meeting, you’re already losing.

Train your own team to articulate the function the same way you do. If your manager describes customer marketing as “running advocacy and the community” and your director describes it as “lifecycle, advocacy, community, education,” your team is already misaligned. Pick one definition. Use the same words. Repetition is the only thing that changes how a function is understood.

Translate every program into the language of the stakeholder you’re talking to. CS hears retention and adoption. Sales hears pipeline influence and references. PMM hears proof and launch enablement. CRO hears revenue impact and CAC payback. Same work, different framing. The work doesn’t change. The translation does.

The pattern under all of this is simple. You are the curator of how the function is understood inside your company. Nobody else is going to do that job for you. The CMO won’t. The CRO won’t. Your peers won’t. The moment you stop educating, the definition starts drifting back to whatever the loudest stakeholder needs from you this quarter.

The one mistake to avoid

The biggest mistake customer marketing leaders make is treating the definition conversation as a one-time exercise. They write the charter, present it once, and assume the org has internalized it. Six months later they’re frustrated that nobody understands the team’s scope, and they take it personally.

It’s not personal, organizations are in constant flux and leaders are context switching 20+ times a day. We need to make it easier for them to remember us which is why I also created this app in claude code as your cheat sheet on what CLG is and how to tell that story. Companies onboard new execs, restructure teams, change priorities, and forget. The function gets re-defined by attrition every quarter. Your job is to keep redefining it back to the truth.

Define it. Then educate forever.

Next week

Issue 19 takes on the salary and structural ceiling that nobody in customer marketing wants to talk about. The data from UserEvidence’s salary study shows 78% of CMA professionals want to advance, but only 12% have a clear path. We’ll walk through what that means for your career, the moves that close the gap, and the move nobody mentions: leaving on time.

Talk soon,

— Kevin