Overview
AI Visibility measures how your brand shows up when people ask AI assistants the kinds of questions your customers actually ask — "best dentist near me," "what does [your brand] do," "is [your brand] any good." More and more people start their search in ChatGPT or a Google AI Overview instead of a list of blue links, and those answers are generated, not indexed — so whether you show up, how you're portrayed, and who you're compared against all depend on what the AI actually says, not just where you rank. AI Visibility gives you evidence instead of guesswork.
What it checks
A check probes seven AI answer engines with a set of real-customer questions and measures your brand's presence in the answers:
- ChatGPT
- Gemini
- Claude
- Perplexity
- Grok
- Google AI Overviews
- Microsoft Copilot
The specific model version behind each engine is managed centrally on the platform side, not hardcoded in the app — so as ChatGPT, Gemini, Claude, Perplexity, and Grok ship newer versions, your checks keep probing a current model without waiting on an app update.
For each question and engine, it looks at whether your brand appeared at all, how prominently, and what tone the answer took — then rolls all of that up into a few headline figures:
- Brand visibility score — the overall percentage of answers you appeared in.
- Sentiment — a −100 to +100 read on how positively AI answers portray you.
- Share of voice — how your mention count stacks up against the named competitors that also came up in the answers.
- Citation sources — where the AI engines say they're getting their information from: your own site, a competitor's, a directory listing, or social media.
- Per-engine breakdown — the same figures split out by engine, so you can see if you're strong in Perplexity but invisible in Copilot, for example.
Measured honestly
Not every engine answers every question every time, and engines occasionally have an outage or a bad response mid-check. When that happens, that particular engine's result for that question is simply left out of your score — it's treated as "not measured," never as "you weren't mentioned." A check that hits a snag along the way still finishes and reports what it did manage to measure, marked as partially complete rather than silently counted as a miss. Your score only ever reflects what was actually observed.
History adds up over time
Every check you run adds to your history rather than replacing it. Each run keeps its own record, and a rollup snapshot is saved alongside it, so you can always look back and see how your visibility, sentiment, and share of voice have moved since the last time you checked — whether that's a week ago or a few months ago. There's no need to run checks on a fixed schedule to get a trend line; every run you (or the assistant) start adds one more point to it.
Where to find it
You start a check by asking the assistant in a session — see Running a check. The results open as a report you can dig into, and the same underlying figures also power four always-on reports that live on every campaign's Insights tab — see Reports. Once a check has run, it also produces a set of prioritized recommendations for closing the gaps it found.
FAQ
Q: Do I need to do anything to start seeing history, or does it run on its own? A: You (or the assistant) need to ask for a check — AI Visibility doesn't probe engines on a background schedule today. Every check you run adds a new point to your trend, so the more often you check in, the more useful the trend becomes.
Q: Does a bad or missing answer from one engine hurt my score? A: No. If an engine can't be reached or fails partway through a check, its result for that question is excluded from your score rather than counted against you.
Q: Do I need to set up anything before running my first check? A: No, but results are noticeably better with a filled-in Business DNA — see Setup notes for what helps most.