Local authority is the connected local record that lets a machine verify a Tampa Bay or St. Petersburg firm before it names that firm to a buyer. It has four moving parts: one place identity, proof that other people provide, answers that name the place, and measurement with the questions local buyers actually ask.
You may know the pattern already. The phone used to ring because people across the bay knew your name. The work is as good as it ever was, and the introductions are thinner. That is not a verdict on your work.
The route to you has changed. A buyer who asks an AI assistant who to call is asking something that only knows what it can check.
This guide takes the four parts in order, shows how each one connects to the next, and says plainly what each asks of you.
Why a Tampa Bay reputation does not reach the machine on its own
A referral does a lot of work in one sentence. One person tells another that you are good, and the trust travels with the sentence: your history, your manner, the time you went past the brief. None of that needs to be written down for it to work.
A machine works differently. It reads a record, and only a record. Put plainly:
AI can't feel your reputation in a room. It can only work with what it can verify.
The closest any platform comes to describing how reputation gets counted is Google's own help page on local results. Google says its local results are "mainly based on relevance, distance, and popularity." In its section headed Prominence, it says prominence "means how well-known a business is," and that it is also based on information such as how many websites link to your business and how many reviews you have.
That is Google's local ranking, on Google's surfaces. It is not a published rule for ChatGPT, Gemini or any other assistant. But it shows the shape of the problem. Reputation starts to count once it leaves a trace that other sources can see.
Now add how buyers ask. Your buyer's AI already knows their situation. It recommends whoever it can verify will look after exactly that person. For a Tampa Bay firm, that turns a vague worry into a practical question: what does your local record let a machine verify about you, here?
Our earlier piece on AI visibility in Tampa Bay looks at which categories in this market are most contested. This one explains the parts that decide whether you can be checked at all.
Part one: one place identity everywhere the machine looks
The first part is the least glamorous. It is also the one most often broken. Every surface that describes your firm has to describe the same firm, in the same place.
Google's guidelines for Business Profiles put it directly. A business should be represented "as it's consistently represented and recognized in the real world across signage, stationery, and other branding." Your name on Google should match the name on your storefront, your website and your stationery. If you travel to clients rather than receiving them, Google asks for one profile for the central office with a designated service area.
Your own website carries the same facts in a form machines read directly. Google's documentation for LocalBusiness structured data lists the business address as a required property. Schema.org defines a property called areaServed as "the geographic area where a service or offered item is provided," and it accepts an administrative area such as a county. So a firm in St. Petersburg that serves clients in Hillsborough and Pasco can say so in markup, not just on a map.
Then there are the directories, aggregators and listings that repeat your details across the web. We build and keep your directory listings consistent. What matters there is agreement. One wrong phone number repeated across listings is still one wrong phone number.
What you do in this part:
- Check that your name, address or service area, phone and category read identically on your profile, your site and your main listings.
- State your service area by county in your markup, matched to where you really work.
This is the work our Entity Identity methodology describes in full.
Part two: proof that other people provide
Next comes evidence that does not come from you. AI recommends businesses it can verify (foundation), that others vouch for (validation), and that know something it doesn't (information gain). We build all three. This part is the vouching.
For local work, the most visible vouching is reviews. Google names review count and positive ratings as things that can help local ranking. Links from other local sources, such as an association you belong to or a publication that covered your work, are the other half of what Google calls prominence.
There is a rule underneath this now, and it is worth knowing. In August 2024 the Federal Trade Commission announced a final rule banning fake reviews and testimonials. It covers reviews that misrepresent that they come from someone who does not exist, which the FTC says includes AI-generated fake reviews. It also bans offering payment or other incentives conditioned on a review expressing a particular sentiment, positive or negative.
So the job is collecting the proof you have already earned. A client who was delighted three months ago is a review that does not exist yet.
What you do in this part:
- Ask recent clients for a review, in their own words, and reply to the ones you get.
- Our advice is stricter than the rule: do not offer anything in return for a review at all.
- Keep a list of the local organizations that would genuinely vouch for you, and make sure they link to the right name.
Our piece on whether Google reviews help in ChatGPT goes deeper on what reviews can and cannot do outside Google.
Part three: answers that name the place the way buyers do
Your own pages are the third part, and they are where most local firms leave the machine guessing.
Start with the name of the market. The federal government treats this region as one metropolitan area, Tampa-St. Petersburg-Clearwater, made of four counties: Hernando, Hillsborough, Pasco and Pinellas. The Census Bureau's own records divide it further into two metropolitan divisions, St. Petersburg-Clearwater-Largo and Tampa. Buyers, meanwhile, say Tampa Bay, or St. Pete, or the name of their neighborhood.
That is four or five names for one market. If your pages use only one of them, or use them inconsistently, a machine has to infer which market you serve. A page that names the place plainly and consistently, in the words your buyers use, removes the guess. Our guide to how ChatGPT chooses local businesses shows how an assistant turns "near me" into a named place before it searches.
The second job is answers. A page that answers a local buyer's question in its first sentence can be lifted whole. A page that opens with your history cannot.
The third job matters most. Your judgment about local situations, the cases you have handled and the mistakes you have watched people make, lives mostly in conversations. We call that the Black Box Effect: genuine expertise that stays invisible because it exists in forms machines cannot find or attribute. It is also the one part of your record no competitor can copy.
What you do in this part:
- Choose the place names you will use and use them the same way everywhere.
- Write one page that answers the question your best local client asked you last month, in the first sentence.
Part four: measure with the questions local buyers ask
The last part turns the other three from a hope into something you can watch. We test the five engines your buyers actually ask: ChatGPT, Gemini, Perplexity, Claude and Grok, with the questions they actually ask.
For a Tampa Bay firm, that means questions with the place in them, written the way a buyer would say them out loud. It means asking them on a schedule, keeping the answers word for word, and noting two things each time: whether you were named, and what was said about you. Good local questions sound like a person, not a keyword list. "Who should I talk to in St. Pete about selling a family business?" is closer to how people talk to an assistant than "business broker Tampa" ever was. Write yours from real calls. The question a client asked before they hired you is the best test question you have.
One dramatic screenshot proves nothing. A sampled set of questions, asked the same way month after month, shows direction.
Here is the honest boundary. Google documents what feeds its own local results. No AI assistant has published how it weighs local listings, reviews or markup when it recommends a business. So none of this work can promise a recommendation. It makes your firm easier to verify, then measures what happens.
What you do in this part:
- Write five local questions in a buyer's own words, with the place named.
- Ask them across engines, and keep the answers, not just your impression of them.
How the four parts work together
Each part feeds the next. Without one place identity, proof attaches to the wrong listing. Without proof, a consistent identity is just a well-kept directory entry. Without local answers, the machine has nothing of yours to repeat. Without measurement, you cannot tell which part is failing.
If you only have time for one part this quarter, start with part one. Nothing else attaches properly until your firm reads as one business in one place. Then collect the proof. Then write.
I wrote in our founder essay, We Asked an AI Where We Should Move, that 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. Local authority is that idea applied to one market.
It is also what our full program runs for firms where place matters. The Answer includes up to four Google Business Profile posts a month where relevant, a citations network built at setup and refreshed quarterly, and monthly five-engine monitoring of named competitors. It is $2,500 setup + $1,500/month, twelve-month program, and the free Recommendation Check tells you which plan your market actually calls for.
If you want the thinking behind every part, our methodologies are published in full. If you want to know where you stand first, see where you stand.