How do I check what ChatGPT says about my business?

Open ChatGPT and ask it the same questions your buyers ask: "who should I hire for [your category] in [your market]?" and "what do you know about [your business name]?" Then do the same in Perplexity and Google AI Overviews. What comes back tells you whether AI names you, ignores you, or sends buyers to a competitor, and each of those outcomes has a specific cause and fix. Here is the full method, prompt by prompt.

Why this check matters more than your rankings

Your Google position tells you where you stand in a list buyers increasingly skip. What the AI says when someone asks for a recommendation is the new front page, and most business owners have never once looked at it. The check takes ten minutes, costs nothing, and regularly changes how founders see their entire marketing. I have watched owners discover their competitor being recommended by name while their own business did not exist to the model.

The method, prompt by prompt

Step 1: the buyer question

Ask the question a real buyer would ask, in their words, not yours. Not your brand name, the problem:

• "Who should I hire for [category] in [city/market]?"
• "Best [service] for [type of customer]?"
• "I need help with [problem your business solves]. Who do you recommend?"

Run each in ChatGPT. Note which businesses get named, in what order, and what the AI says about each. You are checking whether you are in the answer, and who is.

Step 2: the identity question

Now ask about you directly: "What do you know about [business name] in [city]?" and "Is [business name] good?" This checks the second layer: whether the model can build an accurate picture of you at all. Wrong facts, outdated services, or "I don't have information about that business" are all findings.

Step 3: repeat across engines

Run the same prompts in Perplexity, which cites its sources openly, giving you a map of exactly which pages feed the answer, and in Google with AI Overviews on. Different engines read the web differently: ChatGPT's web search leans on Bing, Google AI Overviews pulls from the Google index. Being visible in one and invisible in another is common, and diagnostic.

How to read the three outcomes

Outcome one: you are named

Good. Now check the details: is the description accurate, are the services current, does it position you the way you would? Being cited with wrong facts is its own problem, usually caused by inconsistent information across the web. And check who is named beside you: the AI usually gives two or three names, and your goal is to be the first with the strongest framing.

Outcome two: you are ignored

The most common result, and the most fixable. The model either cannot find you or cannot build a confident case for you. The causes stack in order: AI crawlers blocked in robots.txt, no structured data, no llms.txt, a weak entity, or nothing on the third-party surfaces the AI reads. This is the legibility gap, and it closes with work, not luck.

Outcome three: a competitor owns your answer

Stings, but read it correctly: the AI did not judge them better, it found them easier to read and confirm. Study what Perplexity cites about them. Those citations are your to-do list.

The advanced layer: from snapshot to system

• Test monthly, same prompts. AI answers shift as models update and the web changes. The trend, gaining or losing presence, matters more than any single check.
• Vary the phrasing. Buyers ask the same thing five ways. Test synonyms and adjacent questions; you can be cited for one phrasing and absent for another.
• Log competitor share. Track how often each competitor is named across your prompt set. That is citation share, the core GEO metric, and it turns a feeling into a number.
• Check the logged-out state. Your ChatGPT history biases answers toward you. Test in a fresh session or another account for the neutral view a buyer gets.