What ChatGPT says about your Rhode Island business right now
Here is a test that takes five minutes and costs nothing. Open ChatGPT. Then open Claude, then Gemini or Google's AI Mode, then Perplexity if you have it. In each one, type the question a customer would actually ask, in your town. Not your business name. The question. "Who is a good plumber in Warwick, Rhode Island?" "I need a dentist in Cranston that takes new patients, who should I call?" "Best place to get a dog groomed near East Providence." Then write down what comes back, word for word, and do it again in a private window so your own history is not helping you.
That log, more than any report anyone sells you, is where AI optimization for a Rhode Island business actually starts. It is also the part most owners skip, because they assume the answer is either "we're in there" or "nobody uses that yet." Both are usually wrong, and you can only find out by asking.
What you will probably see
I have run this for a lot of local businesses now, mine included, and the results sort into a handful of piles.
A competitor gets named, confidently, with a phone number. Or the assistant gives three businesses that are all in Providence when you are in Westerly, because it could not find anything closer it trusted. Or it names a business that closed two years ago. Or it names you, with the old phone number and a category you have not used since 2019. Or it hedges, says it cannot recommend anyone specific, and suggests you check Google.
Every one of those is a different problem with a different fix, which is why the verbatim log matters. Writing down "we don't show up" gets you nowhere. Writing down "the engine describes us as a tile installer with a disconnected number" gets you a to-do list.
What the assistant is actually reading
These assistants do not know your business. They read about it, the same way a new customer would if they were thorough and had ten seconds. Your pages, if the text is actually in the HTML. Your Google Business Profile. The directories that list you, whether or not you asked them to. Your reviews, and what the reviews say you do. Local news, the chamber, a vendor who links to you. Then they weigh all of that for consistency. Same name everywhere? Same phone? Same category? When the evidence agrees, they will say your name. When it does not, they say the name of whoever's evidence does.
That is the whole mechanism, and it is why the names that come back are rarely the ones with the best tagline. They are the ones the engine can read and cross-check.
Four reasons the engine names the other guy
1. Their page is readable and yours is a slideshow. Their service page says the plain sentence, "we repair furnaces in Cranston and the surrounding towns," in text an engine can parse. Yours has the same information in an image, or in copy that loads after the page builder's script runs, and a fair number of engines never run that script. From your chair the page looks fine. From the engine's it is blank.
2. Their category is unambiguous. Their profile says "plumber," their site says "plumber," and the structured data under the site says LocalBusiness, plumber, Warwick. Yours says "home solutions" on the profile and "your trusted partner" on the site, and the engine has to guess. It does not guess in your favor.
3. Their entity is consistent and yours is three businesses. The old number is still on two directories. The Yelp listing has the address from before you moved. A business two towns over shares half your name. The engine is not lying. It is reading what is out there and reporting the version it can verify. Entity cleanup is unglamorous and it is the cheapest fix on this list.
4. Other people vouch for them. Reviews that mention the service and the town by name. A listing on the chamber site. A local paper that mentioned them once. Real links from real places. Twenty of those outweigh anything you could add to your own site, and most of the small businesses I look at have a handful at most.
AI optimization for a Rhode Island business: what flips the answer
Fix them in the order above, cheapest first. Clean the entity, which is mostly correcting listings you forgot existed. Then rebuild the service pages so the answers are in the HTML, one service and one geography per page, with the organization and service marked up so the engine can state what you do in machine-readable terms. Then go get the proof, which is the slow part and the part nobody can shortcut.
Expect the Google map pack to move in thirty to sixty days once the profile and listings are right. Expect the AI answers to take longer. They re-read slowly, and the answer changes when the weight of the evidence changes, not on the day you publish something. In my experience it compounds over about ninety days, and the way you know it is working is that you asked the same questions last month and wrote down the answers. That is exactly what we do for clients on the local SEO and AI search program: the same questions to the major engines every month, logged verbatim, and a note the month the answer flips.
Two things that do not work
Putting "Rhode Island's top-rated AI-optimized" anything on your homepage. Engines do not trust taglines, which I wrote about back in Field Notes 02, and a claim with nothing behind it reads to them the same way it reads to you.
Buying a package called "AI SEO" that turns out to be a list of directory submissions and a blog post about ChatGPT. Directory work is part of entity cleanup, and it is worth doing, but it is a chore, not a strategy, and it will not by itself make an engine say your name.
Run the test first
Before you pay anyone, including me, run the five-minute version tonight and keep what comes back. If you are on the Cape, the seasonal version of this problem has its own wrinkles, and AI search optimization on Cape Cod walks through them. If the answers you got are wrong, or missing, or naming someone across the state, the assessment below is the full version of the same read, with the plan attached.
— Kyle Tysvaer, Founder, Insightful Eye Marketing