AI Visibility
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.
Overview
Search does not stop at Google anymore. A growing share of queries goes through AI engines that answer in one synthesis, with a few brands named and a few sources cited. If you are not in that answer, you are invisible to that prospect. AI Visibility measures exactly that: your presence in AI answers, prompt by prompt.
You declare your brand identity (name, variants, domains), the real questions your prospects ask, and your known competitors. The hub queries an AI search engine on each prompt, several times, then detects whether you are named, whether one of your domains is cited as a source, and at which rank. It all aggregates into a score from 0 to 100.
AI answers are non-deterministic: the same prompt can name you one time out of two. So each prompt is probed several times and aggregated into rates, with a stability flag. An unstable signal is shown as uncertain, not as a fact, so the dashboard does not flip from one week to the next.
What the module does
Brand profile
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).
- Canonical name and aliases (an abbreviation, a different spelling)
- Owned domains, to detect a citation as a source
- Market language and scope
- Auto-config: paste your site URL and the hub drafts a profile
Tracked prompts
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.
- Categories: understanding, benefits, comparison, pricing, alternatives
- Assisted generation from the profile, deduplicated against existing prompts
- Each prompt can be toggled on or off without deleting it
Mention, citation, score
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.
- Mention rate and citation rate per prompt
- Average rank of the first mention, a salience proxy
- Stability flag: shown as uncertain when the runs disagree
Competitor leaderboard
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.
Trend and automatic scan
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.
- Twelve-week trend series, global and per prompt
- Weekly automatic scan, cost bounded per run
- Stable prompts probed less often, to save cost
- Manual scan on demand from the hub
How to use it
- 01
Set up your profile
You fill in your name, aliases, domains and market in /seo. Or you paste your site URL to start from a proposed draft.
- 02
Add your prompts
You write the questions your prospects ask AIs, or let the hub generate them from your profile, by intent category.
- 03
Declare your competitors
You list the brands you want to benchmark against, with their variants and domains. They feed the leaderboard.
- 04
Run a scan
Each prompt is probed several times. Detection aggregates mention, citation and rank into a stable result.
- 05
Read your score
The overall 0 to 100 score, the per-prompt detail, and the competitor ranking show where you are visible and where you are missing.
- 06
Track the trend
The weekly scan re-measures automatically. You follow the twelve-week curve and see whether your actions move the needle.