seenby

Explainer

The AI visibility score: what it means and how to read it

A single percentage is a fine headline and a poor guide. Here's what an AI visibility score measures, why 100% can still hide a problem, and the three numbers to read next to it.

Updated November 16, 2026 · 6 min read

Most AI visibility tools lead with one number: the share of answers that name your brand. If you track 25 questions on 3 engines and 30 of the 75 answers name you, your score is 40%. It's simple and comparable week to week. It's also easy to misread.

What's a good score?

It depends almost entirely on which questions you track. Branded or narrow questions produce high scores; broad category questions produce low ones. As a rough guide for a balanced set of category, use-case and alternatives questions:

  • Category leaders: 60–90%
  • Established challengers: 25–60%
  • New or niche brands: under 25%, often concentrated in a few use-case questions

Compare yourself against your own past and your direct rivals on the same questions, never against another company's score from another tool.

Why 100% can hide a problem

A score of 100% means you were named in every answer. It doesn't mean you were recommended first. If a rival is named ahead of you in most answers, buyers read their name first and may stop there. A perfect score on easy questions can also mean your question set is too narrow. If you see 100%, look at:

  1. Position. Average position and how often a rival is named before you.
  2. Harder questions. Add “alternatives to [leader]”, comparison and use-case questions until you find where you're missing.
  3. Engines. A high average can hide one engine where you're absent.

Read it with three other numbers

NumberQuestion it answers
Average positionWhen named, am I the recommendation or a footnote?
Share of voiceAm I gaining or losing against specific rivals?
Cited sourcesWhich pages shape these answers, and am I on them?

Expect noise

AI answers vary from run to run. A move of a few points in one week is often noise; a move that holds for three weeks is real. Track enough questions (20+) and judge trends, not single runs.

Turn it into a to-do list

The useful output isn't the score; it's the list behind it: the questions where you're missing, who is named instead, and the pages those answers cite. That's why Seenby's dashboard opens on that list rather than on the percentage.

Frequently asked questions

How is an AI visibility score calculated?
Answers that name your brand (or cite your site) divided by all answers collected for your tracked questions, usually per engine and overall.
Why did my AI visibility score drop this week?
Answers vary between runs and models update. Check whether the drop holds for more than one run, which engine it came from, and which rivals or sources replaced you.

Track it every week, from $12/month.

ChatGPT, Perplexity, Gemini, Google AI Mode and AI Overviews. Share of voice, cited sources and full answer history.

See plans