# Derek Sato, Founder at Relay Consulting Group — read of conversation-intelligence-ai, May 19 2026

> 9 years running a sales-ops agency for mid-market B2B teams, currently managing 11 clients and looking for something to productize so I stop trading hours for dollars. Two kids, both in competitive swimming, which means my Saturdays are gone from October through March.

## How I got here

My friend Joel, who runs a RevOps consultancy in Austin, dropped this in our group chat last Thursday with zero context. Just the URL. That's the highest-trust referral I get. I bookmarked it and finally opened it on my lunch break today.

## What I clicked first

The headline pulled me in: "Know what was really said in every customer call." That's actually a good headline. It's not "AI-powered insights for your go-to-market journey" garbage. It's a human pain phrased like a human would say it.

But then I realized I had no idea what I was looking at. This isn't a product. It's an idea for a product. I had to reread the page twice to understand the concept. You're not buying conversation intelligence software. You're buying the dossier to BUILD conversation intelligence software yourself. That realization took longer than it should have.

## Where I paused

The financial disclosure section stopped me cold. They show you a "-$38,390 Year-1 take-home" and "1 in 9 Meaningful-success odds" right on the page. I've never seen a vendor just put their Fermi math in the hero like that. My first instinct was skepticism because it felt like a stunt. My second instinct was that no one doing pure marketing theater would publish a negative number that prominently. That tension kept me reading.

## What I distrusted

"Honest disclosure: we don't have live customers on this idea yet." Okay, I respect that sentence. But then the page is called Conversation Intelligence and the score is 72/100 with buyer clarity at 10/10. That 10/10 feels assigned, not earned. Who gave it the 10? The same studio that built the page. There's no outside voice here at all. The whole scoring framework is self-referential. The machine is grading its own homework.

Also: the pricing is $5 to unlock the dossier, $99-$199 to adopt the build. That's a real low price signal for something they describe as a 12-18 week, $48K investment to bring to production. The math doesn't add up on what the $99 actually buys you. "Working code starter, brand assets, copy library, and outreach pack" could mean anything from a GitHub repo with six files to something genuinely useful. I have no way to know.

## What would convince me

One person who bought the $99 package, built the product, and is currently at Month 4. I don't need them to be profitable. I need to see what they actually got, what the code starter looked like, what they had to build themselves on top of it. A raw founder diary post from someone mid-build would do more than five polished case studies.

Also, show me the ICP worksheet from inside the dossier. Just one section, blurred or not. Give me a taste of whether the strategic thinking is sharp or whether it's ChatGPT with a Wishdeal watermark.

## What I'd ask in an email reply

1. The $48K build estimate -- is that the cost if I hire the "operator partnership" team, or is that what I'd spend building it myself on top of the $99 starter? Those are very different situations.
2. What specifically does "working code starter" mean? Is it a Next.js scaffold with a transcription API wired up, or is it closer to a design doc with some pseudocode?
3. Has anyone in your current customer base (even early access, even friends) shipped an MVP from one of these dossiers to actual paying users? Even a single person.

## Verdict: on-the-fence

The honesty mechanics are genuinely unusual and I'm not ready to dismiss that. But I'm buying the thinking, not the product, and I have no evidence yet that the thinking is better than what I'd get from a sharp consultant I already know.

---
*Memo by skeptic persona, generated 2026-05-19. Studio breaks own self-grading loop.*