I promised you a deeper look at the Trust Board I used in the last post. Something I refer to as a “panel of your most important audience members sitting in the room, each quietly asking: does this sound good?” And I said it deserves its own moment in the sun.
Sunglasses on, here it is.
The Trust Board
The Trust Board is a set of six personas I use to evaluate content before it goes anywhere near the site. It’s not an abstract set of demographic buckets, I created a set of actual people with names, backstories, job titles, specific anxieties, and specific things they’re hoping to learn when they read something I wrote.
Why? Well, if you’ve been writing content for any length of time, you know the failure mode that is easy to slip into … you write something, you read it back, you think “this is good” … and then you share it and get a bunch of “WTF is this?”, or you end up trying to rationalize disparate feedback and observations from multiple people. This is at best frustrating, and at worst sould destroying, depending on how excited you are by the topic.
The Trust Board puts other people in the room, as you write, to test thinking and draft options. Yes they are fictional, and yes it’s just AI mimicking and reflecting, but as I’ve honed these, I’ve found that it becomes a pretty good litmus test, as they are by design, on the more testing and poking side of feedback. Aint no sugar coating in my set of personas.
Building the Personas
Here’s where I’ll be honest and say that AI is doing a lot of the work here, because it really is, and in the balance of Humans + AI, this is very slanted to AI. The six personas on my Trust Board were developed in one go as a first iteration based on a scoping I set, and then over time, I refined them when either I saw something that was unhelpful, or based on comments made by people I chat to.
The key thing that I found, was that building useful personas didn’t really require a specifc format or structure. It was the clarity I provided in what role I wanted the AI to use as the baseline and from which to formulate opinions. A vague persona definition, like “mid-career marketer, wants to grow” is basically useless. A persona that gets a bit more granular like “has been in PMM for seven years, is genuinely good at it, and is starting to feel like the execution work is eating her alive while AI tools eat into the edge she thought he had” … that’s a much better definition that can give you real feedback. The difference in those two statements is essentially the emotional box in which the persona lives, and if it was a real person, would inform (cynical me would say taint) their feedback.
Each of my board members has:
- A name and a face in my head (this matters more than you’d think). Some of them are roughly correlated to real people I work and interact with on the regular.
- A specific job situation and company stage, once again based on either real moments in my own life, or that I see in my daily job.
- The exact question they’re showing up hoping to answer. This one I have the benefit of having a lot of life experience, and have lived some of these “pointy questions”.
- The thing that would make them bounce immediately. These were fun to brainstorm, and allowed me to pander to my inner cynic and think about what really annoys me personally as one factor around how these are built.
I’m not going to publish all the details, but I will introduce the band:
- Marcus is an early-stage founder, 18 months into building a product, doing his own marketing and quietly panicking about why his messaging is getting lower than desired engagement.
- Priya is a growth-stage founder post-Series-A who knows she needs a PMM but has never hired one and doesn’t know what “good” looks like. Priya is very much modeled on a real friend.
- Derek is an accidental PMM, technically a product manager who like most PM’s, thinks he can wrote good messaging but he’s only looking at it from the product perspective. He’s tasked with attempting to be a pmm, and missing the underyling science of the art. All us PMM’s know a few Derek’s.
- Simone is six weeks into her first real PMM role after two years in content, smart and eager, drinking from the firehose, and worried about saying something dumb in front of senior stakeholders. Every PMM has been there.
- James is the mid-career PMM who’s genuinely good at his job and is starting to feel like he’s running on a treadmill. Solid execution, unclear upward path, watching AI eat into what used to be his differentiator. I can wave wildly at many real world people.
- Renata leads a team of five PMMs at a medium sized public company, is well respected, stretched thin, and is trying to figure out how to scale the function without redesigning the engine while it’s running. Every PMM team leader just nodded.
Six people. Six very different reasons they’d open something I wrote. Sixteen very different reasons they’d close it.
How I Actually Use It
When I finish a either a piece of a draft or sometimes the whole first draft, I don’t just read it back to myself. I run it through the trust board and ask them to render an opinion on how useful and accurate the content is.
In practice, I run a skill in Claude, and Claude applies each persona’s lens: Would Marcus care about this? Does Priya see herself in it? Would James find this too basic, or would it actually give him something he hasn’t named yet?
The output is a structured review. For each persona: relevance (would they stop scrolling?), resonance (does the framing feel written for them?), actionability (can they do something with this today?), and a verdict. Strong fit, partial fit, not the target for this piece.
That last category is really really important. Not everything needs to land with all six, and is why the blog page has a bunch of topic and audience filters on it. In fact, if I review something and it gets good analysis with all six, I probably wrote something so generic that it didn’t say anything to any of them of value. A piece that’s a strong fit for Simone and Derek and a soft fit for James is helpful. A piece that’s supposed to help Priya hire her first PMM shouldn’t feel like it needs to also resonate with Renata. That’s a different problem in a different moment if time.
What I’m looking for is: did I hit who I was aiming for? And did I accidentally exclude someone I should’ve reached?
What It Catches
The Trust Board hopefully catches 4 things.
The assumed baseline. I’ll write something that presupposes the reader has a certain level pf knowledge about a topic. The Trust Board flags when I’ve inadvertantly applied an overly aspirational expectation. “Marcus would feel talked past here.”
The wrong frame for the right topic. I might write about AI in PMM workflows in a way that’s technically correct but framed around team-level rollout. Simone doesn’t have a team. She’s one person figuring out how to be more effective herself. The topic applies to her, but the frame does not. That’s a rewrite and recast, not really fixable.
The implicit seniority bias. A lot of PMM content (mine included when at times, although this board exists because I’m trying hard not to) defaults to the perspective of someone who’s been around long enough to already have opinions. James and Renata get that. Simone doesn’t yet. Derek is developing his. If I’m writing for Derek and the whole piece assumes confidence he doesn’t have, I’ve written for the wrong version of him.
The thing that’s technically interesting but emotionally irrelevant. James is the one who surfaces this. He’s been around. He’s seen all the various frameworks, written the content, walked the mile. If I write something that’s interesting as a concept but doesn’t connect to the specific thing he’s grinding against right now, the treadmill, the ceiling, whatever analogy you like best, he moves on. He doesn’t have the patience for intellectual exercise without a practical outcome.
What This Has to Do With AI
The Trust Board is fundamentally an AI workflow and I think it’s one of the cleaner examples of Humans + AI I’ve used in practice.
The AI is doing all the heavy simulation work here. I had to describe and build the personas, but from there I run the skill and Claude does the rest. I think that’s genuine positive leverage … doing in minutes something that would otherwise require me to either recruit six real humans to read every draft I write (yeah, that aint happening), imagine six different readers in my head simultaneously (possible but I have no inner monologue, so good luck with that), or just FAFO and see what lands.
But here’s what the AI can’t do for me. It can’t tell me whether the emotional resonance is there in a post. It can’t feel whether a sentence lands with the kind of quiet certainty that makes someone nod and feel they got value. It can flag that Renata might find something too tactical, but whether the piece has the kind of voice that makes Renata trust it in the first place? That’s at best guesswork by Claude. The Trust Board is a pressure-test, not a rendered judgment. The final call on “does this sound like me, and does it do what I need it to do” is still a human call, it’s my call. I re-read these posts many times and futz with them probably one too many times. But that’s kinda fun for me in a Type II way, this whoe sit exists because it brings me joy, and if only 1 human in all of time reads 1 thing on this site and gains value, I win.
It’s the line I keep coming back to as my baseline … AI is a vibe engine that is good at simulating, modeling, and applying a framework consistently across six personas without losing the thread. The content creator, in this case … me … is the source of truth on whether it actually resonates.
Build Your Own
I think every PMM should have something like this. Not necessarily exactly like I have, maybe six personas is too many, and maybe not in this format. But if you’re creating content for an audience, no matter what it is, and you’re the only one reading your drafts before they go out, you’re playing Duke Nukem on hard mode.
Start with an outline of the kind of people you would go to in real life. Capture why you would go to them, and then build those profiles. Give them names. Give them a specific enough job situation that you can hear how they’d react to a sentence. Build them from your real audiences, conversations you’ve had, customers you’ve talked to, people who’ve commented after reading something. The more specific, the more useful.
And then the next time you write something and think “this is good,” run it past them first.
They will tell you things you don’t want to hear. That’s the point.
Adam
PS … I ran the board on this post.
Felt like the only honest thing to do, and there’s something kinda fun about running the board against a post about the board.
This is an abbreviated output of what I see:
**Marcus:** Partial fit. The concept resonates — he's writing stuff alone with no feedback loop — but the post assumes you already have an audience to model personas against. He doesn't quite yet.
**Priya:** Not the target. She's not creating content, she's trying to hire someone to do it. She'd nod and move on.
**Derek:** Strong fit, with a flag. His description in an earlier draft read as a bit of a face-slap ("thinks he can write good messaging but really can't") — not a great look when he's the one reading it. I updated it. The board caught it. (see, told you)
**Simone:** Strong fit. This is probably the persona the post serves best — solo practitioner, no peer network, needs a feedback system she can actually use.
**James:** Partial fit. He's sophisticated enough that the concept isn't new. What holds his attention is the AI angle — specifically the part about what it can't do.
**Renata:** Partial fit, with the most interesting gap. She's thinking about this at team scale, and the post is written entirely from the solo-creator perspective. "The Trust Board for Teams" is a follow-up that probably needs to exist.