What to do when AI says something wrong about your business

A business owner I know found out the hard way. Someone asked ChatGPT to recommend a good auto shop in her area. ChatGPT named her business — then said she closed at 4pm on Saturdays. She closes at 6. She'd had that Saturday schedule for three years.

She had no idea how long the wrong hours had been out there. Or how many people never showed up because of it.

This is more common than most business owners realize. ChatGPT, Google, Perplexity, Gemini — they're all answering questions about your business right now. Some of what they're saying is accurate. Some of it isn't. And unlike a wrong number in the Yellow Pages, there's no one to call and have it fixed by Friday.

Here's what to do about it.

Start by finding out what AI is actually saying

You can't fix what you haven't found. Before anything else, spend 20 minutes running your business name through each of these:

• ChatGPT — Ask: "What can you tell me about [Business Name] in [City, State]?"
• Perplexity — Same question. Perplexity shows its sources, which tells you where the information is coming from.
• Google Gemini — Same question.
• Google Search — Type your business name and look at the info panel that appears on the right side of results.

Write down everything that's wrong. Hours, address, phone number, services, ownership, how the business gets described. Be thorough. The most damaging errors are often the small ones — a wrong closing time, an old address, a service you stopped offering two years ago.

Why it gets it wrong

AI models don't browse the internet the way a person does. They're trained on text pulled from across the web — review sites, directories, news articles, old cached pages — up to a certain point in time. After that cutoff, they stop learning from new sources unless they're connected to live search.

So if your old hours were listed somewhere two years ago and made it into training data, that's what the model learned. It's not fabricating information intentionally — it learned the wrong thing from a source that happened to be wrong or outdated.

The other problem is conflicting information. If your Google Business Profile says one thing, Yelp says another, and your website says a third, the model has to make a judgment call about which source to trust. It doesn't always pick the right one. Google itself will sometimes override your profile details if it believes another source is more accurate — which can quietly change your hours, categories, or contact information without you knowing.

Fix the sources AI reads from

You can't submit a correction directly to ChatGPT's knowledge base. What you can do is fix the sources it reads from — which is the same thing, just on a longer timeline.

Google Business Profile is the highest-priority fix. Google's AI, including the AI Overviews that now appear at the top of search results, reads your Business Profile directly. Log in at business.google.com, verify that every field is current, and set a calendar reminder to check it quarterly. Hours, services, photos, business description — all of it. Wrong hours are one of the fastest ways to lose a customer who was already looking for you.

Your website should state clearly, in plain text: what you do, where you are, your hours, and your phone number. Not buried in an image. Not only in a contact form. Actual readable text on the page that a search engine — or an AI — can find without guessing.

A FAQ section helps more than most people expect. "Where are you located?" "What are your hours?" "Do you offer [specific service]?" AI pulls from Q&A-style content because it mirrors the questions real people ask. Structured, direct answers on your own site are one of the most reliable ways to influence what AI says about you.

Directory listings — Yelp, Apple Maps, Bing Places, BBB, and the other major directories — all feed into AI training data. If your information is inconsistent across them, clean it up. Your name, address, and phone number should be identical everywhere. Same abbreviations, same format, same spelling.

Your Google reviews are part of what AI reads when it summarizes your business. A model that sees 40 reviews mentioning "fast response" and "honest pricing" is going to use those phrases when it describes you. You can't control what customers write, but you can respond — and your responses are part of what gets read too.

What you can't fully control

You can't force an AI model to update immediately. Large language models are retrained periodically, not in real time. Perplexity and Google pull from live sources more frequently than ChatGPT does, so fixes you make today may show up there faster.

Expect corrections to take weeks to months to propagate, depending on the tool. Some platforms let you flag a wrong response directly — there's usually a thumbs-down or "report" option — but there's no guarantee it gets addressed quickly.

What you can control is making sure the authoritative sources are clean, consistent, and current. The AI catches up to the sources. It just takes time.

The one thing most business owners skip

Most of the business owners I talk to aren't ignoring this problem — they just don't know it exists. They've never searched their own business name in ChatGPT. They've never looked at what Perplexity says about them.

That 20-minute check at the top of this article is the thing most people skip. Do it once and you'll know exactly where you stand. If everything looks right, you'll have peace of mind. If something's wrong, you'll have a list — and a list is something you can work through.