Check Your AI Visibility: Does ChatGPT Find Your Business?

Mario Fahlandt · September 28, 2026 · approx. 6 min read

If you want to check your AI visibility, you need neither an agency nor special software. Five targeted questions to ChatGPT, Perplexity and Gemini will show you within minutes whether AI assistants know your business, recommend it, or ignore it completely. This guide walks you through the self-test and then explains which technical criteria determine whether your website is readable for AI systems at all.

Why this matters

More and more buying decisions no longer start at Google but in a chat window. Someone looking for a tax advisor, a metal fabricator or a software product increasingly asks an AI and gets a ready-made list of recommendations back. If your business is not on that list, you simply do not exist for that customer. Unlike Google, there is no page 2 where you might still be found.

The good news: whether a business appears in AI answers is not random. It measurably depends on what is published about the business online and how machine-readable its own website is. Both can be checked and improved.

The 5-minute self-test: 5 prompts to try

Open ChatGPT (ideally with web search enabled), Perplexity or Gemini. Ask the following five questions, adapting industry and location to your business. Important: ask the way a customer would ask, not the way you would search yourself.

Prompt 1: The recommendation question

Recommend a [industry] in [city/region]. Name 3 to 5 providers with a short reason for each.

This is the most important question. Does your business appear? In which position? And who is listed instead? Write down the competitors mentioned.

Prompt 2: The direct question

What do you know about the company [your company name] in [city]?

Here you see whether the AI knows your business at all and which facts it states. Watch for outdated details, wrong services, or mix-ups with similarly named companies.

Prompt 3: The problem question

I have the following problem: [typical customer problem]. Who can help me with this in [region]?

Customers often describe their problem instead of the solution. A roofing company should also be mentioned for "my roof is leaking", not only for "roofer wanted".

Prompt 4: The comparison question

Compare [your company] with [strongest competitor]. What are the strengths and weaknesses?

This question reveals the tone: is your business described positively, neutrally or critically? And which sources does that judgement rely on?

Prompt 5: The source question

Which sources would you use to find information about [your company]?

In Perplexity you see the cited sources directly. Does your own website appear? Or only business directories and review platforms whose content you do not control?

How to interpret the results

• You are recommended and the facts are correct: Good starting position. Now it is about keeping that position and checking regularly, because AI answers change with every model update.
• You are mentioned, but with errors: Action needed. Wrong statements in AI answers almost always come from contradictory or outdated sources. We cover what helps in ChatGPT gets your business wrong.
• You do not appear at all: The most common case for small and medium-sized businesses. The cause is usually a website that AI crawlers struggle to read, combined with too few external mentions.

A note on methodology: a single answer is a snapshot. AI answers vary with phrasing, time of day and model version. Repeat the test with slightly different wording before drawing conclusions.

The technical side: 8 criteria of AI readability

Whether AI systems can use your website as a source comes down to technical details. These are the eight criteria we also check in the Vispra audit, and you can verify each one yourself:

1. robots.txt permissions for AI bots: Does your robots.txt block GPTBot, ClaudeBot or PerplexityBot? Then those systems cannot read your content. Check at your-domain.com/robots.txt.
2. llms.txt: A compact text file that gives AI systems the key facts about your business. Our llms.txt guide shows how.
3. Structured data (JSON-LD): Machine-readable company data in the source code, such as name, address, services and opening hours following schema.org.
4. Title and meta description: Clear, descriptive page titles instead of "Home" or "Welcome".
5. Heading structure: One H1 per page, logical H2/H3 hierarchy. AI systems use headings to understand content.
6. Machine-readable company facts: Name, location and services as text in the HTML, not only as images or hidden in JavaScript.
7. Sufficient text content: A page with three sentences and five photos gives the AI nothing to cite.
8. Answerability: Does your website answer the questions customers ask? FAQ sections and concrete service descriptions beat marketing phrases.

Checklist

• 5 prompts tested in at least 2 AI systems
• Mentioned competitors noted
• Factual errors in AI answers documented
• robots.txt checked for bot blocks
• llms.txt in place or planned
• Structured data present in the source code
• Follow-up test scheduled in 4 weeks

The last point is the one almost everyone forgets. AI visibility is not a one-time state; it changes constantly. If you test once and never again, you will not notice when a model update drops your business from the recommendations.