Is Your Business Showing Up on ChatGPT? How to Check in 2026

Some of your customers are asking ChatGPT, Perplexity, or Google AI who to hire before they ask anyone they know. What those tools say about you shifts by engine, by day, and by how the question was worded. A dated baseline captures what they said at one moment, which is the only version of this you can hold still long enough to work with.

You can run a useful baseline in an afternoon with no paid tools. This walkthrough covers what to record and how far to trust it. The guide to AI search answers for local businesses compares what these engines visibly retrieve and cite, without pretending to know the machinery behind them.

First, what "AI search optimization" actually means

AEO stands for answer engine optimization. The work is measuring what the answer surfaces say about you, then improving the accuracy and reach of the sources they pull from. Nobody can make an engine recommend a business, which is why the measuring half of the job carries as much weight as the fixing half.

Step 1: Choose three relevant answer surfaces to test

Three surfaces are enough, and they will not agree with each other. Product modes, citation display, and the answers themselves differ between them. Record what each one gives you.

ChatGPT. Note which mode you are in and whether web search and citations are switched on before you start. Access and answers vary by plan, account, date, and wording, so the mode is part of the record.

Perplexity. Go to perplexity.ai. It is built around live web search and shows its sources, which makes it the clearest place to see which pages an AI is pulling your information from.

Google's AI results. Run the same search on Google and note whether an AI answer appears, where on the page it sits, and which sources it shows. Availability and presentation keep shifting. Google's official blog documents the features it ships, though never the local-selection logic underneath them.

Step 2: Ask the exact questions a customer would ask

The non-branded questions are the ones that matter here. Searching your own business name tests whether an engine knows you exist. Asking who the best plumber in town is tests whether it will put you forward. Run both sets and keep them apart in your notes.

• "Who is the best [your trade] in [your city]?"
• "Who should I hire for [the specific job] near [your city]?"
• "I need [your service] in [your city], who do you recommend?"
• "What are the top-rated [your trade] companies in [your city]?"

Run each question across all three surfaces. You can also test follow-ups such as "can you give me a few more options" or "who else?" Keep any extra names in a separate column, because a business that shows up only after prompting is telling you something different from one that led the first answer.

"Branded and non-branded questions test different things. Record both, along with the engine, date, location, answer, and citations."

Step 3: Read the answer with three specific questions in mind

When the results come back, check three things in this order.

Are you named at all? Record a yes or a no for every query on every engine. A column of noes is a baseline covering the questions you asked on the day you asked them, and that is all it covers.

Is the information accurate? If you appear, check the phone number, service area, hours, and the sources it cited. Answers repeat stale details from old listings all the time. Fix the underlying source where you can reach it, and log the wrong answer with its date so you can tell later whether the fix took.

Where and how are you mentioned? Record the ordering, the wording used about you, and any links or citations. Being named third in a list is worth logging and worth nothing on its own until you can tie it to a call.

For Corpus Christi service businesses, summer heat is a reason to test emergency-AC questions in July instead of February. Compare whatever the engines say against your tracked calls and bookings for the same weeks.

Four possible gaps to investigate when a business is missing

If the checks came up empty, four gaps are worth investigating first. Query wording, location, product mode, and retrieval availability move the result too, and the platforms keep those to themselves.

1. Check business identity online. Your name, current address or service area, and phone number should read the same everywhere they appear. Two versions of your phone number across four directories is an ambiguity you can remove for free, and worth removing on its own merits whatever an engine makes of it.

2. You have a thin website or no website. A single page of slogans gives a customer almost nothing to read and a retrieval system almost nothing to quote. Publish the plain facts on a page a crawler can reach: services, area, credentials, contact. Whether that page is what kept you out of an answer is a separate question you cannot settle from the outside.

3. Review third-party evidence. Current, genuine reviews do a job your own site cannot do for you, because they are the part of your reputation you did not write. How much weight any engine puts on them stays the platform's business.

4. Check service and location detail. "We handle panel upgrades and EV charger installs across the north side of the city" is a sentence a machine can quote back. "Quality you can trust" gives it nothing. Write the first kind and accept that an engine may still paraphrase it or skip past it.

For local service businesses in Savannah, cold winter nights are a reason to run the same question set on heating work in January. Document every run: one is an anecdote, the same test repeated through the year is worth reading.

What to do first

Start with the factual conflicts you can see. Claim the profiles that apply to you, fill them in properly, and reconcile the name, address or service area, and phone number across every source that carries them. Order the rest of the work by what your own query set turned up, since no generic priority list knows which questions your customers ask.

Next, read your main pages the way an engine would and check that they state services, locations, credentials, and proof in plain sentences. The guide to auditing ChatGPT answers about a local business covers what is observable from outside, and the AEO introduction explains how to keep a baseline current.

For Montgomery home service contractors, inbound movers and storm season both suggest questions worth adding to the list. Add them, run them on a set date, and file the answers with the rest.

Run the check yourself before you hire anyone, including us. It costs nothing, it takes an afternoon, and it leaves you with a dated baseline nobody can talk you out of. I push people toward that order for a self-interested reason as well as an honest one: a business that has never measured its own answers has no way to judge what a visibility report is worth. If you would rather have someone run a broader query set, document the visible gaps, separate out the causes nobody can see, and prioritize the fixes you control, that is what we do.