FOUNDER ESSAY · AUGUST 17, 2026 · 9 MIN READ

We Asked an AI Where We Should Move. Then We Listened.

We fed an AI our real constraints and asked where in the United States we should live. It said St. Petersburg. A second AI agreed. Then we verified it like humans, and moved.

By Christopher Shaw
TL;DR — WHAT TO REMEMBER

We asked an AI where in the United States we should move, it answered St. Petersburg, Florida, and we live here now. That one sentence carries our whole year, so let me unpack it properly.

The Winter That Made Up Our Minds

My wife Jacquie and I were coming to the end of another Texan winter, and Texas was going a little stale for us. We loved a lot about it. But we were landlocked, and being landlocked had started to feel claustrophobic. We had lived most of our lives near water. Add the dryness and the run of 105 degree summer days, and the relief we were craving had a shape: it looked like an ocean.

There was a lifestyle ache underneath the climate one. We used to live in Collingwood, in Melbourne, Australia, where cafes and culture and creative business all sat on the same street. We missed that texture. Austin has a real tech scene and a frenetic energy, and plenty of people thrive on it. We did not. The race to make it big was never our race.

A Research Task, Not a Throwaway Prompt

So we did what we do professionally, except this time the client was us. We fed all of our weird quirks and desires and dreams and expectations and requirements into the AI and asked it where in the United States would be a great place for us to start over.

We did not say "look at Florida." We gave it free rein across the whole country, and we gave it the real inputs, not the tidy ones:

It was a well thought out research task, not just a throwaway prompt. That distinction matters more than anything else in this story. It is exciting to picture us typing "where should we move?" and packing boxes the next morning, but the real version was considered and technical. The quality of the answer came from the quality of the ask.

St. Petersburg came back as the top pick.

The Double-Check

Did we argue with it? No. Did we laugh? Also no. We were asking for advice, and arguing with advice you asked for is counterintuitive. But we did not accept it on faith either.

We took the same parameters, unchanged, and put them into a second AI. It also recommended St. Petersburg. Other places made the shortlist along the way, cities on the East Coast, New Orleans, and we had privately been leaning toward South Florida. The machine kept coming back to the same answer anyway.

Two independent systems, given the same honest inputs, converged on the same city. Here is the thing we tell clients, and we held ourselves to it: model agreement is a signal, never proof. Different assistants still read much of the same internet, so agreement can mean convergent evidence or just a shared source. Two AIs agreeing earns an investigation, not a moving truck. So we started investigating.

Then We Verified It Like Humans

The next phase looked nothing like magic and everything like homework. We followed the links. We watched YouTube walking tours and independent videos by locals. We researched the economy, the chamber of commerce, the support for startups, the attitude toward AI, and the makeup of the business community, right down to costs of living. The city makes that homework easy: St. Petersburg's own economic development pages lay out its business districts and the five target industries it is actively growing, with data analytics named among them. For an AI visibility company, that detail read like a welcome mat. We even asked the AI about the business community we would be joining before we arrived, then checked what it told us against primary sources like those.

It wasn't blind faith. We took the data points and the reasoning and verified ourselves. There was a high level of research, and then a transfer of trust that the research had earned.

Then we came down for a long weekend in June to see it with our own eyes. By late July we had packed the house into a U-Haul, loaded the '80s project Porsche onto a trailer, and convoyed down I-10 through Mobile and New Orleans.

Eight weeks from recommendation to residency. That is what trust looks like when it has been earned in stages: machine reasoning, human verification, then commitment.

What We Found When We Arrived

The first thing that struck me was the water, which is not a surprise. The second thing was harder to name. St. Pete has a more relaxed version of cool, not the showy, uptight version we had gotten used to. Jacquie put her finger on it before I could: in Texas everyone was busy, but it never felt like they were creating. Here, she said, you see people who are creating, not just performingly being busy.

The coffee scene, and I say this as someone who takes coffee unreasonably seriously, was draw card number two after the water. The arts felt proactive rather than decorative. This is a city that hangs its art on the outside of its buildings: the annual SHINE festival, run by the St. Petersburg Arts Alliance, has put nearly 200 murals on the walls downtown since 2015. And where other cities talk endlessly about building community, Jacquie's read was that here community is a byproduct of activity. People make things, and the belonging follows.

It's a mini Melbourne, I said on our second day. Different suburbs, different vibes, water at the end of the street. It felt like home faster than any place we have landed.

We Are the Buyers We Describe

Here is why I am telling you this story on a company website. We run Probably Genius, an AI visibility company. Our whole thesis is that a class of researched buyers now uses AI for guided discovery: they hand an assistant their real constraints, receive a reasoned recommendation, verify it, and then spend serious money. We wrote the portrait of the Researched Buyer months before we became the case study. We chose a city, and everything that comes with relocating a family and a company, exactly the way we describe buyers choosing an accountant or an architect.

And the guided discovery did not stop at the city limits. The same AI-assisted conversation helped us think through how to position Probably Genius inside St. Pete's business community: where the crowded ground was, and where the open ground was. It steered us away from planting our flag beside every SEO voice in town and toward the seat nobody had claimed, AI search and referral. An assistant that can advise a business on where it fits in a city can certainly advise a buyer on which business in that city to call.

The difference between this and Google is the difference between discovery and impulse. A search result, or a cute Instagram post of St. Pete Beach, can spark interest. It cannot walk you from your own quirks and constraints through evidence to a decision. Guided discovery can, and that is what the Recommendation Layer now does for buyers every day.

There's real money being spent by the suggestions of AI. Houses bought, cars purchased, advice needed and paid for. And when the AI sends the buyer, the buyer arrives decided: a Seer Interactive case study measured ChatGPT referrals converting at 15.9 percent against 1.76 percent for Google organic on the same site. Ahrefs found AI search was half a percent of its traffic but drove 12 percent of its product signups. A case study and one company's books, not a market average, and that is the point: buyers who arrive from an AI conversation did their considering inside it. Fewer visitors, far more buyers. Our move sits on the other end of the same behavior, one household and one company deciding where years of spending will land, guided by an AI they questioned hard and then trusted.

What This Means If You Run a St. Pete Business

Our own preliminary research keeps showing St. Petersburg being recommended by AI, for a whole variety of reasons. There is enough growth and activity and energy here for the machines to have noticed that something special is happening. Buyers like us are already asking.

The question is whether AI can verify, understand, and recommend you when they do. You cannot control what an AI assistant says. You can control whether your expertise is structured, on the record, and checkable when the question is asked. That is readiness, and it is the entire job. It is also, openly, the methodology we publish and run for our clients.

If you want to know where you stand, start where our clients start: the Recommendation Check, our free 109-Point AI Visibility Diagnostic, hand-scored across eight zones and now run from St. Petersburg, Florida. It shows you what AI can currently verify about your business, and what is missing.

We asked an AI where to move. Then we listened, we verified, and we chose. Your next client is doing the same thing right now. Probably about you.

Frequently asked questions

Which AI picked St. Petersburg?
We deliberately do not name it. We used one mainstream AI assistant for the original research task and a second, entirely separate one to check the answer, and both recommended St. Petersburg. The vendor is the least important part of the story. The method is the part worth copying: honest inputs, free rein, a second independent opinion, then human verification. Any capable assistant, given a serious research task instead of a throwaway prompt, can run this play.
Did the AI make the decision for you?
No, and that distinction is the whole point. The AI did the discovery: it took our real constraints and reasoned its way to a recommendation we would not have reached on our own, since St. Petersburg was not on our private shortlist. We did the verification: independent research, primary sources, and a scouting trip. Trust transferred in stages, and the machine earned each one. That is exactly how we watch researched buyers choose accountants, architects, and advisors.
What should a St. Pete business do about buyers like us?
Make yourself verifiable. You cannot control what an AI assistant says about you, but you can control whether your expertise is structured, on the record, and checkable when a buyer's assistant comes asking. Start by finding out what AI can currently verify about your business: that is what the Recommendation Check exists to show you, at no cost. It is the same first step every client of ours takes, and it tells you plainly where you stand before you spend a dollar fixing anything.
How do we get started with Probably Genius?
Start where every client starts: the Recommendation Check, our free 109-Point AI Visibility Diagnostic. It shows you what AI can currently verify about your business and what is missing, before you spend a dollar. From there, our most complete published program is The Answer: $2,500 setup and $1,500 a month across a twelve-month program, or $2,000 a month with no minimum term, with twelve done-for-you thought-leadership articles planned to create approximately 60 clear Knowledge Entries every month. You run the business. We make sure AI knows why to recommend it.

CITATIONS

  1. "Case Study: 6 Learnings, 1 Site — How Traffic from ChatGPT Converts" (Seer Interactive). Single-site B2B case study, October 2024 to April 2025: ChatGPT referral traffic converted at 15.9 percent against 1.76 percent for Google organic on the same site. seerinteractive.com
  2. "AI search traffic conversions" (Ahrefs, first-party data). AI search accounted for half a percent of Ahrefs' site traffic and drove 12.1 percent of product signups. ahrefs.com
  3. "About the Districts" (City of St. Petersburg Economic Development). The city's own pages on its business districts and the five target industries it is actively growing, with data analytics named among them. stpete.org
  4. "SHINE Mural Festival" (St. Petersburg Arts Alliance). Confirms the annual festival has created nearly 200 murals across downtown St. Petersburg and its arts districts since 2015. stpeteartsalliance.org
WRITTEN BYChristopher Shaw

Co-Founder and Chief of Ops at Probably Genius, an AI visibility firm helping expert-led businesses become the named answer in AI search. A background in award-winning music production and years directing creative strategy and digital operations, now spent turning complex expertise into records machines can verify. Let's talk →

Related reading

Read the method.
Then see it run on you.

The 109-point diagnostic is this library, applied to your business. Free, about an hour to present, no obligation.

See where you stand →

A real diagnostic, not a sales call in disguise.