AI Visibility

Why AI gets your business wrong

Ask ChatGPT about your business and the answer often has an old price, wrong menu or a service you dropped years ago. It isn't random. There are five common reasons, and most trace back to one thing: AI can't find a comprehensive, current, reliable source about your business, so it fills the gaps with guesses.

9:41
●●●
Does Marlow Cafe do catering?

From what I can find, Marlow Cafe looks like a small dine-in spot — I don't see any mention of catering or private events. You may want to check with them directly, or look at other cafes in the area that list catering.

Ask anything
What AI got wrongMarlow has catered events for years — a whole revenue line AI can't see, so it's steering the customer to competitors.

Supporting every major AI assistant

ChatGPT
Claude
Gemini
Perplexity
Copilot
Grok
The short answer
AI doesn't fact-check your business. It retrieves what it can find, then repeats it, gaps and all.
Most AI crawlers can't see parts of your website: so menus, prices and availability inside widgets, images or PDFs can be invisible to them.
When a fact is missing, models often guess a confident answer rather than say "I don't know."
One accurate, current source of truth for key information leaves far less room for stale or conflicting facts to slip in.
The five reasons

Where the wrong answer actually comes from

Four of these five are gaps: old data, conflicting sources, unreadable formats, or nothing at all. The fifth is what AI does when it hits a gap.

1. It's remembering an old you

A model's core knowledge is frozen at a training date. Old prices, a number you dropped, a location you closed can all linger.

2. The web disagrees about you

Your site says one thing, an old directory says another. When sources conflict, AI may merge them or pick the wrong one.

3. Your best info is unreadable

Most AI crawlers can't see parts of your website. Menus or pricing loaded by a widget, a booking app, or a PDF often aren't seen.

4. It guesses to fill the gap

When a fact is missing, models are trained to produce a confident, plausible answer rather than admit they don't know.

5. Even AI doing live search is often wrong

A Columbia Journalism Review study of AI search tools found answers were incorrect more than 60% of the time, often citing sources that don't exist.

The common thread

No up-to-date, comprehensive, AI-readable source that AI can rely on. Give it one and most of these problems shrink.

93%
of businesses had at least one basic fact — hours, a service, even their name — wrong or missing in an AI assistant's answer. Small businesses get it worst.
Searchable study of 13,000+ AI answers about businesses, via SME Magazine
The fix

Give AI one source it can trust

Most of these problems are the same problem: AI is working from old data, conflicting listings, formats it can't read, or nothing at all, and it guesses. Hand it one accurate, current, AI-readable source about your business and you shrink the room to guess or merge the wrong facts.

That's what Courtyard maintains for you: your services, images, pricing, and policies, published in a form ChatGPT, Gemini, Claude, and Perplexity can read, and kept current as things change. As your business evolves, you can easily tell AI what's changed.

Honest about the limits: a clean source improves accuracy and fills gaps. It can't force an assistant to cite you, and it won't rewrite every old listing on the web overnight. It gives AI the right answer to find.

See how Courtyard works
Before
Old prices, wrong services, guessed details
After
Current prices, right services, accurate
FAQ

Why AI gets it wrong

The questions owners ask when they spot a mistake.

Why does ChatGPT show my old hours or a price I changed?
Its core knowledge is frozen at a training date, and old directory listings still float around the web. So it repeats a stale version of you unless it can find something current and reliable.
My website is correct. Why does AI still get it wrong?
If your hours or menu load through a widget, a booking app, an image or PDF, most AI crawlers can't read them.
Does AI check whether what it says about me is true?
No. It retrieves, synthesizes, and repeats, it doesn't verify. When a fact is missing, it may confidently make one up rather than say it doesn't know.
Where is AI even getting the wrong information?
Usually a mix of outdated listings and conflicting sources across the web. When they disagree, AI picks one or blends them, and sometimes it's wrong.
If I fix my information, will AI instantly get it right?
It gets much more accurate, but not instantly and not always with a citation. A single clean source of truth reduces the guessing; it can't override every old signal on the web at once.

Give AI the information it needs, so it doesn't have to guess about your business

Give every assistant one current source of truth. Set up your AI knowledge base in minutes. From $20/month.

Get started free