What counts as a mention?

Your brand named in the body of the answer the assistant gives. A link in a source list is a citation, which Refine tracks separately: the two move independently, and the gap between them is often the most useful signal you have.

Why does my visibility change day to day?

AI answers are not deterministic: the same prompt produces different answers between runs, and engines re-rank continuously as their sources change. That is exactly why Refine runs prompts daily rather than once: a single check cannot tell a real shift from normal variance.

How many prompts should I track?

Start with the ten to twenty questions that decide a deal in your category, then widen. Tracking hundreds of low-intent prompts produces a number that moves without telling you what to do about it.

AI Search Visibility Tracking

Track how often and prominently your brand appears in AI-generated responses across all major platforms.

Run a free audit See pricing

When a buyer asks an assistant which tools to consider, they act on the handful of names that come back. Visibility tracking measures whether yours is one of them, not whether you rank on a results page nobody opened, but whether you are named inside the answer itself.

Mention rate

The share of tracked prompts where your brand is named in the answer. Ask an assistant the same buying question ten times and it will not name the same brands every time: answers drift between runs. A mention rate of 40% means you were named in four runs out of ten, which is a very different position from being named once by chance.

How Refine measures it

Prompts, not keywords

You track the questions your buyers actually ask, "best CRM for a 20-person sales team", rather than keywords. Each one runs daily against every engine on your plan.

Read the way a buyer reads

Refine reads the AI products themselves rather than querying the model APIs. An API returns a different configuration, without the retrieval and browsing layered on top, so a brand can appear in an API response and never in ChatGPT.

Daily, because one run proves nothing

A single check tells you what happened once. Running daily turns a noisy signal into a trend you can act on, and shows movement as it happens rather than at the end of a quarter.

What teams do with it

Find the prompts you are absent from

Sort by mention rate and the bottom of the list is your backlog: the buying questions where competitors are named and you are not.

Prove that published work moved something

Publish against a gap, then watch whether the mention rate for those prompts rises over the following weeks. It is the closest thing to attribution this channel offers.

Catch drops before they cost you

Engines re-rank as sources change. A prompt you owned can quietly stop naming you, and the daily refresh surfaces it while it is still fixable.

Where this metric misleads

Mention rate alone flatters you if the mentions are unfavourable or buried at the end of an answer. Read it alongside position and sentiment, being named last, grudgingly, is not the same result as being named first.

Key Benefits

Real-time monitoring across 6+ AI platforms

Historical trend analysis

Competitor visibility benchmarking

Custom alert thresholds

Who Uses Visibility Tracking?

Brand managers

SEO teams

Marketing executives

How It Works

Connect your brand

Set up your brand profile and connect your domain to Refine AI.

Configure tracking

Enable Visibility Tracking for your selected AI platforms.

View insights

Access real-time visibility tracking data on your dashboard.

Take action

Use the insights to optimize your AI search presence.

Visibility Tracking: frequently asked questions

What counts as a mention?: Your brand named in the body of the answer the assistant gives. A link in a source list is a citation, which Refine tracks separately: the two move independently, and the gap between them is often the most useful signal you have.
Why does my visibility change day to day?: AI answers are not deterministic: the same prompt produces different answers between runs, and engines re-rank continuously as their sources change. That is exactly why Refine runs prompts daily rather than once: a single check cannot tell a real shift from normal variance.
How many prompts should I track?: Start with the ten to twenty questions that decide a deal in your category, then widen. Tracking hundreds of low-intent prompts produces a number that moves without telling you what to do about it.