How to Check What ChatGPT Says About Your Business

**Open a fresh chat in ChatGPT, Perplexity, and Gemini, ask the questions your customers would ask, then ask each one where it got its information. The sources it names are your fix list. The whole check takes about twenty minutes and most business owners have never done it once.**

Customers are asking AI assistants about local businesses, and those assistants are answering — sometimes accurately, sometimes with information from a directory that has your address from two moves ago. You can't manage what you haven't looked at.

Why Should You Check This?

**Because the answers may be wrong.** Old address, disconnected phone, hours from before you changed them, services you no longer offer. A confident wrong answer sends customers somewhere you aren't.

**Because you may not be named at all.** If an assistant recommends three businesses in your category and neighborhood and you're not among them, that's a visibility gap you can work on.

**Because the sources are the actionable part.** These systems will usually tell you what they read. That turns a vague concern into a specific list of pages to fix.

How Do You Run the Check?

**Use a fresh session every time.** In ChatGPT, use a temporary chat. In Perplexity and Gemini, start new. Otherwise your own history influences the answer and you're testing yourself rather than what a stranger sees.

**Don't sign in as yourself if you can avoid it**, or use a browser where you're not logged in, so personalization doesn't skew things.

**Ask questions the way a customer would.** Conversational, constrained, with a real situation attached. Not keyword phrases.

What Should You Ask?

Run a fixed set every month so the results are comparable. Six is enough to start:

**1. The direct question.** *"What can you tell me about [Your Business Name] in Seattle?"*

Checks whether they know you exist and whether the basics are right.

**2. The category question.** *"Who's the best [your service] in [your neighborhood], Seattle?"*

Checks whether you're recommended at all.

**3. The situation question.** *"I'm in [neighborhood] and [realistic problem]. Who should I call?"*

The constrained, conversational shape people actually use.

**4. The comparison question.** *"How does [Your Business] compare to other [category] in Seattle?"*

Reveals what it thinks distinguishes you, if anything.

**5. The practical question.** *"What are [Your Business]'s hours and where are they located?"*

The accuracy check.

**6. The price question.** *"What does [your service] cost in Seattle?"*

Checks whether you're a source for the pricing question, which is one of the most common things asked about any local service.

The Follow-Up That Matters Most

After every answer, ask: **"Where did you get that information?"**

Most assistants will list their sources. That list is the whole point of the exercise.

You'll typically see some combination of your Google Business Profile, directory listings, review sites, roundup articles, and sometimes your own website. Occasionally something you've never heard of.

Three things to do with that list:

**Fix anything wrong.** If a directory has your old address, go correct it.

**Note who else is cited.** If a roundup article recommends three competitors and not you, that article is a target.

**Notice what's missing.** If your own website never appears as a source, that's a signal about how readable and specific your site actually is.

What Should You Record?

A simple spreadsheet, one row per prompt, one column per month:

- The prompt

- Which assistant

- Were you named? (yes/no)

- Was the information accurate?

- Sources cited

Six prompts across three assistants is eighteen data points a month. Enough to see a trend.

How Should You Read the Results?

**Expect variation.** These systems are non-deterministic. You'll appear in one check and not the next with nothing having changed. This is normal and it is not a sign the work failed.

**Watch the trend across months, not any single check.** A business named in three of eighteen prompts in January and eleven of eighteen in July is winning, even if any given week looks mixed.

**Separate accuracy from visibility.** They're different problems with different fixes. Wrong information means inconsistent sources — a citation cleanup job. Not being named at all means insufficient corroboration — a reviews, content, and third-party mention job.

What Do the Common Results Mean?

**"I don't have information about that business."** You have very little presence in the sources these systems retrieve. Start with the Google Business Profile and citation consistency.

**Accurate but brief.** They know you exist but have nothing specific to say. You need more substance — detailed reviews, described services, documented projects, third-party mentions.

**Wrong details.** Somewhere in your citation footprint the old information persists. Search your old phone number and old address to find it.

**Competitors named, you're not.** They have stronger corroboration — usually more substantive reviews, more third-party mentions, or presence in the roundup articles being cited.

**Named with a reason.** The good outcome. Note what reason it gave, because that tells you which of your signals is landing.

What Do You Do With What You Find?

The fixes are mostly the same regardless of what you find, which is convenient:

- Complete and correct the Google Business Profile, with services listed individually

- Make business information identical everywhere it appears

- Build reviews that describe specific services in specific neighborhoods

- Publish pages that answer real customer questions in plain language

- Get into the roundups and local lists that are already being cited

- Publish your own specifics — pricing shape, process, service areas — so there's something to quote

None of this is exotic. It's the same local visibility work, aimed at a reader that happens to be a language model.

Can You Guarantee an AI Will Recommend You?

No. Nobody controls what these systems say, and any provider claiming otherwise is describing something they can't deliver.

What can be done is make yours the business with the most complete, consistent, and specifically described presence across the sources these systems read. That's the whole lever, and it's a real one.

Key Takeaways

- Use a fresh, logged-out session — otherwise you're testing your own history, not what a customer sees.

- Ask six fixed customer-shaped questions monthly across ChatGPT, Perplexity, and Gemini.

- Always follow up with "where did you get that information?" — the sources are the actionable output.

- Expect week-to-week variation; read the trend across months, and separate accuracy problems from visibility problems.

Frequently Asked Questions

How do I find out what ChatGPT says about my business?

Open a temporary or fresh chat, ideally logged out, and ask the questions a customer would ask: what it knows about your business, who it recommends in your category and neighborhood, and your hours and location. Then ask where it got the information.

Why does ChatGPT have wrong information about my business?

Because it assembles answers from sources across the web — directory listings, review sites, your profile, and articles — and if any of those carry outdated details, they can surface. Searching your old phone number and old address usually finds the listings still carrying stale data.

How often should I check my AI search visibility?

Monthly. Run the same fixed set of prompts across ChatGPT, Perplexity, and Gemini in fresh sessions and log whether you appear and what's cited. Results fluctuate week to week, so monthly checks across several months show the trend without reacting to noise.

Can I control what AI says about my business?

Not directly, and no provider can guarantee an AI will recommend you. What you can influence is the source material — a complete and accurate Google Business Profile, consistent information across directories, substantive reviews, specific content on your own site, and mentions in third-party articles.

Want the Check Run for You?

Watt Consulting runs monthly AI visibility audits for Seattle businesses across ChatGPT, Perplexity, and Gemini, with a source-by-source fix list. Start your project.