How much do AI assistants agree about Dutch brands? Barely
Every quarter we publish what our measurement pool shows about how ChatGPT, Gemini and Claude describe Dutch markets. All figures come from real, repeated measurements; the method is always alongside them.
Source overlap between assistants is almost zero. Visibility at ChatGPT says almost nothing about Gemini or Claude; whoever measures on one assistant measures a third of the playing field.
The largest measured difference in this edition. Assistants build their answers on different sources and arrive at fundamentally different recommendations.
On about half of the questions not a single tracked brand is mentioned. There the first mention is the cheapest: no one has to be displaced.
A handful of platforms (including comparison sites and trade media) feed all assistants at once. Whoever is on them counts everywhere; those are the most expensive and most valuable spots.
What this means if you run a brand
One: never measure on a single assistant; you then see a third of the playing field and draw conclusions about the rest. Two: the answer is determined by the sources assistants cite, and those sources differ per assistant; so your action list is different per assistant. Three: half the terrain is unoccupied, and unoccupied terrain is cheap.
Methodology
The figures come from Geony's shared measurement pool: brand-free customer questions, each asked several times to ChatGPT, Gemini and Claude, with deterministic source extraction from the provider metadata. Edition 1 covers hundreds of measurements across Dutch service markets. Percentages internally carry a 95% interval; the full method is here. Next edition: autumn 2026, with a quarter-on-quarter comparison.
Benchmarks per industry
Where the measurement pool is deep enough, we show an aggregated, anonymized visibility figure per industry on the industry pages. As long as an industry has too few measurements we deliberately show nothing there: a figure resting on a handful of brands would conceal more than it reveals. The series grows along with the measurement pool.
Within a day you'll know whether AI mentions or ignores your brand
The first measurement runs on your own prompts, across three assistants, with real figures. No demo data, no sales call first.
Prefer to talk it over first? Book a call