GEO audit
Visibility audit
The baseline of your AI visibility: how often you are mentioned, cited and recommended, against named competitors. And what is wrong.
A reply within one working day, from the person who measures.
- ChatGPT
- Perplexity
- Google AI Overviews
- Copilot
- Gemini
Which suppliers automate packaging lines in Germany?
Answer: For automated packaging lines in Germany, Hollweg Antriebstechnik and Marenholt Systeme are the names most often given. For smaller runs and special machinery, Rieckert & Sohn is also an option.
Your companynot named
Your company
7 %
Rank 4 of 5
- Hollweg Antriebstechnik38 %
- Marenholt Systeme29 %
- Rieckert & Sohn19 %
- Your company7 %
- Tessmar Industrie4 %
Sample data with invented names. Shows the format, not a client result.
First place on Google is the ticket, not the seat.
AI engines do not simply summarise the top ten. Which sources they prefer in your category is an empirical question. And being mentioned, cited and recommended are three different things.
44%
of pages cited in Google AI Overviews are not in the top 20 organic results for the same query.
33%
of B2B buyers have bought from a previously unknown supplier on an AI recommendation.
Source: G2, 1,076 B2B decision-makers, March 2026 (G2 runs a review platform)
How it runs
The proposal names the timeline once we know your category.
Collect the questions
A conversation with you and your sales team. Questions naming your brand go into a separate group.
OutputPrompt set, frozen
Set competitors
Your list plus the names the engines volunteer. Rewritten per language, not translated.
OutputCompetitor and language plan
Measure
Several runs across five engines, clean sessions, documented region. Every answer kept verbatim.
OutputRaw observations
Analyse
Share per engine, split into mentions, citations, recommendations. Errors, sources, spread.
OutputFindings with spread
Prioritise
A sorted list with effort, impact and evidence. Reviewed with you.
OutputReport and action list
Two readings. What really happened in between.
The bar shows the sample's spread. If the movement lies within it, we measured no change, and the report says so. The third row is the case nobody else shows: up, but within the spread.
| Brand | Baseline | Follow-up | Spread | Verdict |
|---|---|---|---|---|
| Your company | 7 % | 19 % | ±4 | +12, measured change |
| Hollweg Antriebstechnik | 38 % | 31 % | ±5 | -7, measured change |
| Marenholt Systeme | 29 % | 33 % | ±6 | +4, within the spread |
| Rieckert & Sohn | 19 % | 18 % | ±3 | -1, within the spread |
Sample data with invented names. Shows the format, not a client result.
What the report contains
- Share of answers
Per engine and overall, against competitors, with sample size and spread.
- Mentions, citations, recommendations
Three separate figures, because they have three different causes.
- Errors
Every false statement with wording, date and frequency.
- Sources
Which websites the engines cite in your category, and whether yours is among them.
- Technical findings
Crawler access, readability without JavaScript, the first characters of every page.
- Raw data
Date, prompt-set version, runs, engines, wording. Every claim is traceable.
Questions about the audit
Why not a single snapshot?
Because it proves nothing. Only comparing two readings shows whether a figure is stable and whether an action worked.
How many questions are in the set?
As many as your buying situations yield. The number is in the proposal. What matters more is separating questions with and without your name.
Do you analyse competitors too?
Yes, in the same set with the same figures. A technical review of other companies' websites is not included.
Which language do you measure in?
The language your buyers use. German questions surface different brands and sources than English ones. For further markets, questions are rewritten, not translated.
What can you not measure?
We have no view into training data: we see that an engine prefers a source, not why. A published question set stands in for what individual customers actually asked. There is no cleanly isolated revenue contribution, because many AI visits arrive without a referrer.
Start with the baseline.
Tell us your category and one question your customers ask.
A reply within one working day, from the person who measures.