Introduction

People increasingly ask an AI assistant instead of running a search. The assistant answers in prose, names a handful of brands, and the rest of the market is simply absent. AgentMention measures whether your brand is among the ones it names.

What AgentMention does

You define the questions your customers ask. AgentMention runs them against AI assistants on a schedule and turns the answers into metrics:

• Are you mentioned? — Visibility
• How much of the conversation is yours? — Share of Voice
• How early are you named? — Position
• How favourably are you described? — Sentiment
• Which sources do assistants read? — Citations & Sources
• Which questions do you lose entirely? — Prompt gaps

Alongside the answer data, it audits your own site the way an AI crawler sees it — structured data, technical readability and crawler access — and reports which AI crawlers actually visit you.

Who it is for

Marketing, SEO and content teams who already track organic search and now need the same visibility for AI answers. The product assumes you can publish or influence content — that is the lever every metric responds to.

How to use these docs

• New account? Start with the Quickstart, then walk through Onboarding.
• Number on the dashboard you do not recognise? Every metric has its own page under Metrics, each with the exact formula.
• Not sure what a term means? The Glossary is short and definitional.
• Something looks wrong? FAQ & troubleshooting covers the usual causes.

A note on how to read the numbers

Two habits make the whole product easier to trust:

1. Check the scope before the number. Every metric is scoped to one brand, one prompt group, one set of assistants and one period. Two contradictory-looking numbers are usually two different scopes.
2. Read trends, not single values. AI answers vary between runs. A single day is a sample; a fortnight is a signal.

Core concepts covers the vocabulary the rest of the documentation uses.