Method · June 5, 2026 · 6 min read

How to measure your AI visibility (share of voice in AI answers)

You measure AI visibility by running a fixed set of buyer prompts against each AI engine on a schedule, and recording whether your brand is mentioned, cited as a source, and how you compare to competitors. The output is three numbers you can track over time: a visibility score, a citation rate, and share of voice.

The three metrics that matter

Mention rate, of your target prompts, how often is your brand named at all?
Citation rate, how often are you listed as a source (the stronger signal)?
Share of voice (SOV), your mentions/citations as a share of the total across you and your competitors.

Track these per engine (ChatGPT, Perplexity, Gemini, Google AI Overviews) and combined.

How to measure it (step by step)

Define your prompt set. 10–50 questions a real buyer would ask, "best [category] for [use case]", "alternatives to [competitor]", "is [you] good for [job]".
Pick your competitors. The 3–5 brands you expect to compete with for citations.
Run the prompts on each engine. Manually to start, or via API for scale and consistency.
Record the result. For each prompt × engine: were you mentioned? cited? at what position? who else was cited?
Roll it up. Compute mention rate, citation rate, and SOV. Save a dated snapshot.
Repeat on a schedule. Weekly or biweekly, AI answers shift, so a single check is a snapshot, not a trend.

Why "do it once" isn't enough

AI answers vary by phrasing, change as models update, and drift as the web changes. A one-time check tells you where you stand today; only repeated measurement tells you whether your work is moving the needle. The trend is the point.

A useful first-party signal

If you have access to Bing Webmaster Tools, its AI Performance report shows real citation counts from Microsoft Copilot and partners, a rare piece of first-party AI-citation data. Use it alongside your prompt-based measurement.

From measurement to action

Measurement only matters if it drives work. Each prompt where a competitor is cited and you aren't is a target: build or fix the answer page using the AEO checklist, then re-measure.