Mesure si les moteurs de recherche IA citent ta marque, et face à quels concurrents.
When a prospect asks an AI search engine, does your brand show up in the answer? AI Visibility tracks your prompts, probes the engine, and gives you a presence score, citation detection, and a ranking of your competitors.
Free plan. 60 seconds. No CC required.
You configure what counts as "you" in an AI answer. The canonical name, accepted variants, and your domains. Domains are used to detect a citation as a source, not just a name mention. You set the market language and its scope (local, national, international).
A tracked prompt is a real question a prospect asks an AI, one you want to show up on. Each prompt carries an intent category, which tells you where you are strong or absent. You can write them by hand, or let the hub generate them from your profile and market.
For each prompt, the hub separates two things. Mention: your name appears in the answer. Citation: one of your domains appears as a source. Being cited as a source weighs more than being simply named. The overall score (0 to 100) combines the mention rate (weight 0.6) and the citation rate (weight 0.4), averaged over all your prompts.
Question asked to an AIWhich tool should a freelancer use to run their business?
Position 2 of 5 cited brands. Name mentioned and domain used as a source.
Question asked to an AIHow do I invoice a client abroad?
Position 1 of 4 cited brands. Name mentioned and domain used as a source.
Question asked to an AIWhat is an alternative to a classic CRM for a solo worker?
Position 8 of 9. The name shows up at the end of the answer, with no link to the site.
On every scan, the hub picks up the known competitors named in the answers. The leaderboard ranks them by how many prompts they appear in. You see at a glance who captures AI visibility on your questions, and where you are getting beaten.
Measurements accumulate over twelve weeks, as a weekly series, with a per-prompt breakdown. A scan runs automatically: each workspace is re-probed at most once a week, with a per-run cap to bound the cost. A prompt whose result is stable is only re-probed beyond three weeks, since a stable result does not move. You can also trigger a manual scan, which runs in the background.
Free plan, no card. Test the module in real conditions.
Free plan. 60 seconds. No CC required.