How to measure your AI share of voice against competitors

In traditional search you measured yourself by rankings. In AI answers there’s no ranking - the engine just names a few brands and moves on. So the question gets sharper: of all the answers where your category comes up, how often is it you the AI names, versus your competitors? That’s your share of voice, and it’s the most honest single scoreboard for AI visibility.

This guide covers how to baseline share of voice, read it correctly, and move it - using Visibility AI’s Monitor layer.

Why share of voice beats a raw mention count

Presence tells you whether you’re in the room. Share of voice tells you how much of the room is yours. The difference matters because AI answers are a zero-sum space: an engine names two or three brands, not twenty, so every competitor mention is one that could have been yours.

A raw count of your own mentions can climb while your position erodes - if three competitors are being named more often than you, you’re losing the category even as your absolute number rises. Share of voice is relative by construction, which is exactly what makes it a scoreboard rather than a vanity metric. It converts a vague worry (“are we invisible to AI?”) into a defensible position you can watch and improve. There’s more background in our explainer on measuring share of voice in AI answers.

Step 1: Fix your prompt set and your competitor set

Two lists define the metric. In Visibility AI you set both up once.

• The prompt set: the buyer questions you want to be the answer to, phrased as full questions and kept stable so week-to-week numbers stay comparable.
• The competitor set: the brands you’re measuring share against. Include your obvious rivals, but also let the platform surface the brands that keep appearing in your answers - some of your real AI competitors won’t be the ones you’d have named.

Getting these two lists right is 80% of the work. A sloppy prompt set or a missing competitor makes every downstream number misleading.

Step 2: Run the set across all seven engines

Share of voice on one engine is share of voice in one seventh of the market. The same question yields different brand line-ups on ChatGPT, Perplexity, Gemini, Google AI Overviews, Google AI Mode, Copilot and Claude, because each retrieves from a different mix of sources.

Visibility AI runs your prompts across all seven and records, for each answer, every brand named and whether yours was among them. That gives you a share-of-voice figure per engine and one overall - and the per-engine split is often the most actionable part, because it shows exactly where you’re being out-cited.

Step 3: Read it in context, not as a lone number

“30% share of voice” means nothing on its own. Read it three ways:

1. Against competitors. Are you first, third, or nowhere in your category’s line-up? A 30% share that leads the field is a strong position; the same 30% behind a 55% leader is a warning.
2. By engine. Strong on Perplexity, absent on Gemini is a common and fixable pattern - it usually points at which sources each engine trusts.
3. By prompt type. You might dominate comparison questions but disappear on “best X for Y” questions. That tells you where the gap is.

Visibility AI breaks the figure down along all three axes so you’re never staring at a single percentage with no way to act on it.

Step 4: Track the trend and set alerts

A baseline is a photo; share of voice is a film. Because the platform re-runs your set on a schedule, you can watch your share climb after you ship content, or slip when a competitor lands a placement in a source the engines trust. Treat a sustained drop on one engine as a signal to investigate, not noise - it usually means the source picture shifted underneath you.

Step 5: Close the gap where you’re losing it

When share of voice is low, the next question is why those competitors and not you? That’s a diagnostic job. Visibility AI’s Diagnose layer works backwards from the answers to the sources behind them, so you can see the exact review sites, guides and threads the AI leaned on to name a competitor over you - the ground covered in the sources AI engines trust.

From there the moves are specific: earn placement in the sources you’re missing with the Backlink Agent, fix the pages that should be cited with content optimization, and get into the discussions AI reads. Share of voice is how you’ll know whether those moves worked.

See your share today

The quickest way to get a first read is a free Visibility Check; to turn it into a scoreboard you watch, set your prompt and competitor sets in Monitor. Once share of voice is baselined, you have a single number that tells you, honestly, whether you’re winning or losing the AI-answer category - and a trend line that shows whether your work is moving it.

Frequently asked questions

What is share of voice in AI answers?

It's the percentage of relevant AI answers in which your brand is mentioned or recommended, measured against the competitors that appear alongside you. If your category comes up in 40 tracked answers and you're named in 12, your share of voice is 30% - meaningful only in context of who else showed up.

Why not just track my own mentions?

Your own mention count can rise while your position gets worse, if competitors are growing faster. Share of voice is relative by design, so it tells you whether you're winning or losing ground - not just whether you're present.

How often does share of voice change?

It moves as models update, as competitors publish, and as the sources engines trust shift. Weekly tracking catches real movement without chasing daily noise; a single reading is only a baseline.