More and more customers ask an AI before they contact a provider. If you don't appear in the answers, you lose inquiries without ever noticing. Whether that's happening to you can be found out without specialist software.
Why a manual test
is enough to start.
AI answers fluctuate, so single tools quoting percentage scores deserve some caution. A simple, repeatable test with real customer questions shows directly what a prospective customer gets to see.
What matters is discipline: always the same questions, always in a fresh chat window, and a log. Only that lets you tell later whether anything has improved.
The test in
five steps
- 01
Write five typical customer questions
Phrase them the way your customers would, for example "Which web design agency is there in Dessau?" or "Who optimizes Google profiles in the Anhalt region?" Don't put company names in the question.
- 02
Ask each question in at least three systems
ChatGPT, Gemini, and Perplexity draw on different sources. Open a new chat window for each question so earlier answers don't influence the result.
- 03
Log the results
Note whether your business is named, where, with what details, and which competitors show up. A simple spreadsheet is enough.
- 04
Check for accuracy
If address, services, or hours are stated incorrectly, that's a clear sign of inconsistent or outdated data on the web.
- 05
Repeat monthly
A single test is a snapshot. Repeated on the same day each month with the same questions, it shows whether your measures are working.
Bottom line:
Measure first, then optimize.
The test doesn't deliver exact numbers, but it's an honest assessment. If you're not named, you at least know where to start: unambiguous business data, structured data, and content that answers questions clearly.
If you're unsure where to begin after the test, our GEO & AI Search Optimization service shows how we close the gaps systematically.
FAQ
Frequently asked
questions.
Because AI systems take earlier messages in the same chat into account. A fresh window keeps earlier questions or answers from skewing the result and makes the test comparable.
It's an indication, not a verdict. Common causes are inconsistent business data, few current pieces of content, or missing structured data. Since AI answers change, repeat the test several times before drawing conclusions.
About the author
Owner of simnify in Dessau-Roßlau. Builds high-performance websites and leads local SEO and GEO strategy for businesses across the Anhalt region and all of Germany.
