And for new readers, welcome to Signal // Noise, the newsletter read by founders, CEOs, and scrappy builders every Thursday. Each week, what I'm listening to, one deep dive, notes from the field, three links worth your time. No buzz or bullshit.

🎧 Want the whole vibe? Find & bookmark the running playlist right here.

I ran an AI visibility audit on myself this weekend and the result was more interesting than I expected.

I typed in buyer-style prompts to multiple AI tools. "Best coach for bootstrapped founders." "Who are the best advisors for bootstrapped CEOs?" "CEO coach for founders doing $5M to $20M."

I did not show up. Not once.

What came back instead: Dave Bailey, Jerry Colonna, Matt Mochary, Strategic Coach, and a handful of others. Fine company. None of it was me 😢.

Then I typed my name plus "CEO coaching," and the machines lit up. LinkedIn sourced me as the growth-stage CEO's coach for bootstrapped founders doing $5M to $20M. A podcast page with my Hampton and Hustle background. A profile page describing exactly who I work with. An interview or two. A Rebooting episode.

Below is a fresh, no-memory search answer for "Who is Jordan DiPietro?"

Not bad, to be honest!

Because I've only been doing formally coaching founders for a few years, no, I'm not shocked that when it came to recommendations, I'm not the first thing out of ChatGPT's mouth… that would be, to put it mildly, unexpected.

But the shape of the result is what matters, and I'd bet money it could be the same shape as yours.

Here's an important distinction.

AI-visible means the machines know you exist and understand what you do. Entity recognition. If someone types your name, the answer is accurate.

AI-recommended means you show up when somebody who's never heard of you describes their problem.

And those are completely different problems with completely different fixes, and most companies are solving the first one when they have the second one.

Claude scored me, roughly:

Nine and two. That’s the gap I need to solve for.

Because most people who need you aren't typing your name.

They're typing their problem.

Seriously, stop reading and go do it. Open ChatGPT, Perplexity, Claude, and Gemini and run two sets of prompts.

Write down who shows up in set two. That's your real competitive set when it comes to AEO/GEO.

Then, score yourself out of 10 on both. The gap between those two numbers is your actual problem.

Ahrefs studied 75,000 brands and correlated which factors predict showing up in an AI answer:

Mentions beat backlinks roughly 3 to 1.

And the cliff is steep. Top-quartile brands for web mentions averaged 169 AI mentions. The quartile below them averaged 14.

Now the number that should change your calendar: about 85% of brand mentions inside AI answers come from third-party domains, not your own site. Roughly 48% come from community and UGC platforms. And about 60% of AI Overview citations come from URLs that don't rank in the traditional top 20 at all.

Historically, rankings were gated by domain authority, and that's a game you lose to the big boys by default, no matter how good your writing is. You were never gonna win that, especially if you were part of big industries like finance, software, insurance, etc.

But… mentions aren't gated by anything.

My domain rating sucks, for example (the domain is < 1-year old). But, I still show up cleanly for branded queries, because a handful of podcasts and profile pages did the work my website couldn't.

You can't outrank the big guys. You can absolutely out-mention them inside your niche. And with barely a third of Google searches still sending a click, that's the first shift in search in twenty years that structurally favors the small company.

Three things you can actually try to do:

And since most measurement advice assumes volume you don't have, these five work small:

Listen, I'm a 2 out of 10 on the generic stuff, and I'm okay with that for now.

You might be worse, or better; doesn't really matter.

What matters is that the fix isn't gated behind ten years of domain authority anymore, it's gated behind how many other people's pages say your name next to the problem you solve.

That's something you can start on this afternoon, for free, without a single new blog post. So go run the two prompt sets. If you come back a 9 on branded and a 2 on generic like I did, you don't have a content problem. You've got a mention problem, and that one you can actually fix this quarter.

Nothing in this 2024 talk is new, but Patreon's Jack Conte delivers it like a boss. It reframed my scoreboard: the metric isn't subscribers or open rate, it's how many people took a meaningful next step in the last 90 days. Replied, forwarded, bought, booked, introduced you to someone.

Clifford Sosin's argument is that the takeoff never came because we misread what the limit was. Thinking was never the bottleneck. Reality is, and reality answers at its own pace no matter how smart the thing asking is.

Six years post-pandemic, The Journal ran the receipts on remote work and the number that stuck with me is 25%. Not a phase, not a fight anymore, just the shape of the thing now. Worth 20 minutes if you're still building a hiring process that assumes otherwise.

That's it for this week.

And until next time, thanks for reading.

Jordan