Buying comparison · August 2026
Comparing AI visibility audits: four criteria that matter
The market for AI visibility research is young and crowded. Our measurement of 29 August 2026 surfaced 82 Dutch providers alone. This page sets out the four points on which you can compare them fairly, and the data showing why the first one weighs most.
Compare on four points: how many AI platforms are actually queried, which measurement levels are recorded, whether the measurement is reproducible, and whether pricing is one-time or a subscription. The number of platforms weighs most: six platforms used 319 sources between them to answer eight buying questions, and 81% of those sources appeared at exactly one platform.
Our own figures, so you can place them beside any provider
Ask every provider for the same fields. If one of them cannot fill in a field, that tells you something.
| Criterion | Scan €29 | Pro €79 | Complete €149 |
| AI platforms queried | 6 | 6 | 6 |
| Which platforms | ChatGPT, Gemini, Perplexity, Claude, Grok, Meta AI | same six | same six |
| Buying questions | 8 | 20 | 40 |
| Direct platform measurements | 48 | 120 | 240 |
| Measurement levels | retrieval, citation, mention, recommendation | same four | same four |
| Competitors in the report | those that surfaced in the answers | 3, chosen in advance | 5, chosen in advance |
| Subscription required | no | no | no |
| Delivery | usually 20 minutes | about an hour | about 2 to 3 hours |
| Method published | yes | yes | yes |
| Repeatable with identical questions | yes | yes | yes |
The four criteria
11. How many platforms are actually queried?
This is the point with the greatest impact. Measuring only ChatGPT shows 52 of 319 sources: 16%. The usual three — ChatGPT, Gemini and Claude — cover 55%. The remaining 45% sits with Perplexity, Grok and Meta AI, which almost nobody measures.
22. Which measurement levels are recorded?
Being retrieved, cited, mentioned and recommended are four different outcomes. An audit that records only whether you are mentioned misses exactly where the difference lies.
33. Is the measurement reproducible?
Ask whether the same questions can be put again later and whether the method is published. Without repetition you cannot tell whether a change came from your work or from a different moment. Also ask how temporarily unavailable measurements are handled: counted as zero, or reported separately.
44. One-time or subscription?
Pricing models range from single reports to monthly subscriptions and consultancy retainers. Work out what a first picture costs, not what a year costs.
Frequently asked questions
- How many AI platforms should a good audit measure?
- As many as your customers actually use. In our measurement 81% of the 319 sources appeared at exactly one of six platforms, so measuring a single platform shows between 3% and 32% of the picture depending on which one.
- What are the four measurement levels?
- Retrieval: your page was fetched. Citation: it was used as a source. Mention: your brand appears in the answer. Recommendation: you are actively recommended.
- What does an AI visibility audit cost?
- Prices vary widely, from single reports costing tens of euros to projects costing hundreds and monthly subscriptions. Our own Scan is €29 one-time, Pro €79 and Complete €149, all including Dutch VAT.
About these figures. Source-overlap data comes from one blind measurement on 29 August 2026 across six platforms and eight buying questions, 38 of 48 measurements successful. This page deliberately contains no prices or scores for other providers: that information changes and we do not measure third-party websites. Compare for yourself, using the four criteria above.
AI Visibility Report Group · Dutch Chamber of Commerce 42136638 · VAT NL004345918B24 · Terms · Contact