Skip to main content
FullReach AI puts a project’s prompts to the AI platforms on a schedule and keeps every answer. It counts every number it reports from those stored answers. This manual says how. FullReach AI asks each platform through its own app, as a buyer does, and not through the model’s API. So every answer in a counted day is one a buyer could see. How answers are collected says more, and describes the one exception: a new project’s first hour. New here? Get started goes from signup to the first results.

A short introduction to FullReach AI.

​
The unit: an answer

An answer is what one platform returned to one prompt in one market on one day. Every metric is a count over answers, or over the brand mentions found inside them.

​
The four headline numbers

Visibility

The share of analyzed answers that name this brand.

Share of Voice

This brand’s mentions against the mentions of every monitored brand.

Sentiment

How FullReach AI classifies each mention of this brand: positive, neutral or negative.

Position

Where answers place this brand among all the brands they name, counted from 1.

​
Find your way

The manual follows the work, in order.

Set up your project

Which prompts and competitors to track, and who asks in which market.

Interpret your results

Every dashboard page, and how to read each figure on it.

Optimize

The worklist of where to act, with the evidence.

Agent Analytics

The crawlers and the visitors that AI sends to this brand’s site.

Agencies

One project per client, one plan, and reports under your name.

MCP server

Ask an AI assistant about the figures.
Dashboard map lists every entry in the dashboard with its page.