Guide
A guide to AI visibility
The definitions and reading rules used throughout the report.
What is AI visibility?
AI visibility is the share of AI-generated answers in which a brand appears, either named in the answer text or cited as a source. It is measured by sending a fixed set of questions (prompts) to engines such as ChatGPT, Google AI Mode, Google AI Overviews, Gemini, Perplexity, Claude and Copilot, and recording what comes back.
The concept grew out of search engine optimisation, but the unit is different. A search result page lists ten links in a stable order; an AI answer is a short piece of prose that may recommend two brands, cite five pages and change when the same question is asked again. Visibility in that setting is a probability, not a rank, and it has to be estimated from many answers rather than read off a single page.
Why does AI visibility need its own report?
Because the numbers that describe it are scattered, defined differently and often quoted without context. Assistant vendors publish user counts on their own schedules and with their own definitions: weekly users, monthly users, app users, queries. Analytics firms measure referral traffic from panels or client samples. Tool vendors announce funding rounds and customer counts in press releases. Each figure is useful; side by side, without definitions, they mislead.
This report collects those figures in one place, keeps the definition attached to every number and links each one to the source that first published it. Where two sources disagree, both are shown and the difference is explained. Where a number has not been published, the report says so instead of estimating it.
How should you read an AI visibility figure?
Four questions settle most confusion:
- What is being counted? Users, sessions, prompts, answers and citations are different units. A weekly active user figure and a monthly active user figure cannot be compared directly.
- Who measured it? A vendor reporting its own usage, a panel-based traffic estimate and an academic sample answer different questions, with different error.
- When? Products in this category change monthly. Every figure in the report carries the date it was published and the date it was checked.
- Which engine and which surface? Gemini, Google AI Mode and AI Overviews are separate products, even though all three belong to Google. ChatGPT with and without web search behaves differently too.
For brand measurement, the core metrics are the mention rate (the share of answers that name the brand), the citation rate (the share that link to the brand's own domain), the share of voice (the brand's mentions relative to competitors across the same prompts) and the position at which the brand is first named. Vendors define these differently; the chapters note where.
What does this report cover?
The report is organised in chapters. The lead chapter maps the market: how widely AI search is used, who sells AI visibility software, how those companies are funded and how their pricing is structured. A statistics chapter collects usage, referral-traffic and click-through data by engine. Shorter chapters follow the product changes that move visibility most directly, such as the rollout of Google AI Mode, advertising inside ChatGPT and OpenAI's business profiles.
Each chapter opens with a short summary of key findings, ends with a numbered source list and is also available as plain Markdown for readers and machines that prefer it.
What does the report leave out?
Market-size forecasts without a stated method, vendor-reported customer counts that cannot be traced to a primary statement, and any rating or score of products. The report describes the market; it does not rank it. Where it discusses AI visibility software, it lists features and published prices with the date they were checked, and links to each vendor so readers can confirm them.