You ask ChatGPT to recommend suitable service providers, products, or software solutions. It mentions several competitors, but your own business is missing.
The obvious conclusion is:
“ChatGPT doesn’t know my business.”
However, a single answer is not enough to support this conclusion. Your business might appear the next time you ask. The question may not match your positioning. Alternatively, ChatGPT may genuinely lack the information it needs to categorise and recommend your services.
The short answer is:
If ChatGPT fails to mention your business once, this may simply reflect normal variation. If your business remains absent across several relevant questions and over a longer period, it may indicate a genuine visibility problem.
The challenge is to distinguish between these two cases.
Does ChatGPT Know My Business and Recommend It?
It is helpful to distinguish between four levels of AI visibility:
| Level | What it means |
|---|---|
| Awareness | ChatGPT can describe your business when you ask about it by name. |
| Mention | Your business appears in an answer even though its name was not included in the question. |
| Recommendation | ChatGPT presents your business as a suitable solution for a specific need. |
| Source citation | Your website or another page about your business is linked as a source. |
A simple test such as “What do you know about Example Ltd?” does not show whether potential customers would discover the business.
A more commercially relevant question might be:
“Which IT service providers in Potsdam help medium-sized companies implement Microsoft 365?”
If your business appears without being named in the question, it is visible for this specific need. These unprompted mentions are more relevant to customer acquisition than simple recognition of the company name.
Why a Single ChatGPT Test Proves Very Little
Achtung.app submits every tracked question twice. The wording, model, and technical settings remain unchanged. There is only about one second between the two requests.
It then compares which brands appear in both answers.
An analysis of 3,313 answer pairs, collected between 1 June and 18 July 2026, revealed considerable differences:
| AI platform | Average overlap | Exactly the same brands | No brand in common |
|---|---|---|---|
| ChatGPT | 69.0% | 37.0% | 10.0% |
| Perplexity | 48.5% | 9.2% | 8.4% |
| Gemini | 40.7% | 9.7% | 24.3% |
ChatGPT produced the most consistent brand lists in this study. Nevertheless, only 37% of the answer pairs contained exactly the same brands.
In one out of every ten ChatGPT answer pairs, the two responses did not have a single brand in common. A business could therefore be recommended in the first answer and disappear completely just one second later.
Would you like to test several questions yourself? Run a free AI visibility scan.
The study was conducted through the providers’ APIs using fixed models. Consumer applications may behave differently. The analysis also does not determine which recommendation was more accurate or useful. It demonstrates how unreliable a single mention or omission can be.
Why Do the Answers Vary?
ChatGPT does not work like a business directory containing a fixed list of suitable companies.
When a question requires current information, ChatGPT can search the web. The original question may be converted into one or more search queries. Sources are then selected and combined into an answer. The conversation context and, for local searches, the user’s location may also affect the result. OpenAI explains how ChatGPT Search works.
A different selection of sources may be enough to produce a different list. An AI answer is therefore closer to a snapshot than a fixed ranking position.
This does not mean that every poor result can be dismissed as chance. If your business is never mentioned across many relevant questions and repeated tests, you should investigate the possible causes.
7 Reasons Why ChatGPT Does Not Mention Your Business
OpenAI does not publish a complete checklist for business mentions. The following reasons are therefore not guaranteed ranking factors. They provide a practical framework for investigating poor AI visibility.
1. You Asked Only One Question
The most common misinterpretation begins with the measurement itself.
A managing director asks ChatGPT once for suitable providers. Their own business is absent, so they conclude that its AI visibility is poor. A repeated request, however, could produce a different result.
The reverse is also true. A single recommendation does not prove that a business has consistently strong visibility.
Instead of evaluating one answer, test:
- several repetitions of the same question
- different ways of expressing the same need
- various relevant use cases
- multiple AI platforms
- changes over a longer period
Only when your business is consistently absent across this range of tests can you identify a reliable visibility problem.
2. The Question Does Not Match Your Actual Services
An AI system may know your business but still conclude that it does not fit the question.
An IT provider that mainly serves larger medium-sized companies may not appear in response to a request for “cheap computer support for home users”. This is not necessarily a visibility problem. It may be a sensible distinction.
Problems arise when businesses test questions that promise high search volumes but do not match their services, target customers, or service area.
Useful test questions should reflect genuine purchasing decisions:
- What problem leads customers to hire you?
- Which alternatives do they compare?
- Which requirements influence the decision?
- In which regions do you offer the service?
- What size of business is your service suitable for?
The more precisely the decision is described, the more meaningful the result becomes.
3. Your Website Is Not Accessible to ChatGPT Search
A website must be technically accessible before it can appear as a current source in ChatGPT Search.
OpenAI uses OAI-SearchBot for this purpose. If the bot is blocked by robots.txt, a firewall, the hosting provider, or an additional security service, the website cannot normally be included in ChatGPT Search answers. OpenAI therefore recommends allowing the bot and its published IP address ranges.
There is an important limitation: technical accessibility does not guarantee a mention. It merely creates the conditions under which your website can be considered as a current source.
Businesses should check:
- Is OAI-SearchBot allowed to crawl the important pages?
- Are bots being blocked by a firewall, CDN, or bot-protection service?
- Can the content be accessed without signing in or interacting with the page?
- Do important pages return a successful HTTP status?
- Is essential information available as readable text?
Technical accessibility is one of the few factors that can be assessed relatively clearly.
4. Your Website Does Not Clearly Explain Who Your Business Is Suitable For
Many websites list their services but fail to connect them to a specific customer need.
A website might say:
“We develop innovative and individual digital solutions.
This gives little indication of whether the company builds online shops, automates manufacturing processes, or develops software for local authorities.
Verifiable statements are more useful:
“We develop custom web applications for medium-sized manufacturing companies and integrate them with existing ERP systems.”
This description connects several pieces of information:
- the service offered
- the target audience
- the use case
- the technical or industry specialisation
ChatGPT should not have to infer these connections from advertising slogans. They should be stated explicitly on service pages, case-study pages, and company profiles.
5. Important Information Is Inconsistent or Out of Date
Your website may describe your services differently from an industry directory. An old address may still appear in company profiles. Older pages may promote services that you no longer provide.
Such inconsistencies make it harder to categorise your business correctly.
Particularly relevant information includes:
- company name and spelling
- location and service area
- services offered
- target audiences and industries
- contact people
- opening hours and contact details
- current product names
- company profiles and directory listings
This does not mean publishing the same marketing copy everywhere. The underlying facts should, however, be consistent.
6. There Is No Independent Evidence for Your Claims
A company’s own website can claim that it is a market leader, highly experienced, or particularly reliable. Verifiable evidence is more useful for a recommendation than self-promotion.
This evidence may include:
- specific client projects
- detailed case studies
- customer reviews
- expert articles based on first-hand experience
- listings in relevant industry directories
- mentions by associations, media outlets, or partners
- certifications and verifiable qualifications
Not every mention is valuable. A paid listing on an arbitrary website does not automatically create a convincing reputation.
What matters is whether independent sources confirm the connection between your business, a service, and a specific use case.
7. Your Content Answers Company Questions Instead of Customer Questions
Many websites are structured from an internal perspective:
- About us
- Our philosophy
- Our services
- Our solutions
Potential customers express their needs differently:
- Which CRM software is suitable for a company with 50 employees?
- How much does implementing Microsoft 365 cost?
- Which agency has experience with accessible government portals?
- How can a legacy ERP system be connected to an online shop?
If your website does not answer these questions, publicly accessible information connecting your business with the relevant purchasing decision may be missing.
Generic advice articles are only of limited value. An article such as “Ten Benefits of Digital Transformation” could have been published by almost any competitor.
Information derived from your actual work is more valuable:
- typical costs and timescales
- common mistakes
- anonymised project data
- specific processes
- before-and-after comparisons
- well-reasoned recommendations
- the limitations of your own services
AI can help structure this information and improve the language. The underlying business knowledge must still come from your company.
How to Measure Your AI Visibility Properly
A reliable visibility test starts with the questions potential customers ask, not with your company name.
1. Collect Relevant Decision-Making Questions
Use questions from sales conversations, emails, support requests, and traditional keyword research. Do not include your brand name in the test questions.
2. Test Several Phrasings
“Which CRM systems are suitable for small businesses?” may produce different results from “Recommend a GDPR-compliant CRM system for a German company with 30 employees.”
Both questions concern the same topic but describe different purchasing situations.
3. Repeat Every Question
A single answer depends too heavily on how the response was assembled at that moment. Only repeated tests show whether a brand appears regularly or only occasionally.
4. Measure More Than the Mention Itself
Record:
- Is the business mentioned?
- How strongly is it recommended?
- In what context does it appear?
- Which competitors are mentioned more frequently?
- Which sources are used?
- Is the information about your business accurate?
A neutral mention at the bottom of a long list should not be treated in the same way as an explicit recommendation with a relevant explanation.
5. Monitor Changes Over Several Weeks
After improving content or removing technical barriers, do not compare only one new answer with one previous result.
Monitor whether the frequency of mentions changes across multiple questions and points in time. This is the only way to distinguish a genuine trend from normal variation.
Check Whether AI Assistants Recommend Your Business
Achtung.app asks ChatGPT and another AI provider up to five questions about your industry without mentioning your company name. The results show whether your business is mentioned, which competitors appear instead, and where visibility gaps may exist.
The free scan provides a more meaningful sample than one manual question. However, ongoing measurement is still necessary to identify reliable developments over time.
Can a ChatGPT Recommendation Be Guaranteed?
No.
OpenAI states that ChatGPT Search rankings are based on several factors intended to provide relevant and reliable results. A top position cannot be guaranteed. Technical accessibility, clear information, and independent evidence can improve your starting position, but they cannot force a recommendation.
Traditional SEO cannot guarantee the top position on Google either. AI visibility should likewise be understood as an ongoing process, not as a one-time registration with a platform.
Conclusion: Measure Systematically Instead of Asking Once
If ChatGPT does not mention your business, there are two fundamentally different explanations:
- You are seeing normal variation in a single answer.
- ChatGPT cannot find enough relevant and reliable information to select your business for that question.
One test cannot distinguish between these cases.
The assessment becomes meaningful only when it covers several relevant questions, phrasings, repetitions, and points in time. If your business remains consistently invisible, you can then investigate technical accessibility, unclear positioning, inconsistent information, or missing evidence.
This is exactly why Achtung.app does not simply test whether an AI system recognises your company name when prompted. Its visibility check measures whether ChatGPT, Gemini, and Perplexity mention your business in response to brand-free questions and which competitors appear instead.
For customer acquisition, the important question is not whether ChatGPT has heard of your business.
What matters is whether it recommends your business when a potential customer asks for a suitable solution.
