Public measurements
What AIs recommend, measured in public
Our observatories survey several assistants about the same category or topic. Every result is based on real, dated responses and a public methodology.
Which database do AI assistants recommend
PostgreSQL stands out
Which CRM do AI assistants recommend
HubSpot stands out
Do AI assistants recommend themselves?
Tight field, no front-runner
In other languages
Quel CRM l'IA recommande
HubSpot stands out
Présidentielle 2027
Tight field, no front-runner
How to read these observatories
Each observatory asks the same set of questions, phrased the way a real user would write them, to several families of AI assistants. The figures are mention rates computed from their real, dated answers: they describe what AI assistants put forward at a given moment, not the quality of the products or public opinion. A measurement has limits: answers vary from one session to the next and from one model to another, which is why we publish the method, the panel and the dates.
What about when the question is about your brand?
Use the same method to see where AIs recommend you, ignore you or get facts about you wrong.