Getting started
What is Analyze AI?
Analyze AI tracks how AI assistants like ChatGPT, Claude, Gemini, and Perplexity present your brand — and gives you the tools to improve it.
The problem it solves
Your next customer probably won't find you through a list of blue links. They'll ask ChatGPT, Perplexity, or Gemini — "what's the best platform for X?" — and the answer they get decides which brands make their shortlist. That answer layer is invisible from traditional analytics: you can't see what AI says about you, why competitors get recommended, or which sources the engines trust.
Analyze AI makes that layer visible and improvable.
How it works
Track the questions that matter
You define prompts — the questions your buyers ask AI assistants. Analyze AI suggests starter prompts grounded in real search demand and your own brand profile. See setting up your prompts.
Analyze AI runs them across six AI engines
ChatGPT, Claude, Gemini, Perplexity, Grok, and Microsoft Copilot — on a recurring schedule, so you get trends instead of anecdotes.
Every answer is measured
Visibility (are you mentioned, and at what rank), sentiment (how positively AI talks about you), position against competitors, and citations (which pages and domains the engines used as sources — including who they name on every cited page).
You act on what you learn
Perception analysis shows the narrative behind the numbers; citation intelligence shows which sources to win; the content tools write and optimize pages built to earn AI citations.
What's in the platform
| Area | What it does |
|---|---|
| Overview | Your daily read: visibility, sentiment, and position trends against competitors, top sources, and the prompts where rivals get cited but you don't |
| Prompts | The portfolio of buyer questions you track — with unique suggestions grounded in search demand and your Brand Vault |
| Perception | How AI frames every brand in your space: the visibility × narrative quadrant, themes, and competitor battlecards |
| Sources & citations | The pages and domains AI engines trust — with entity tracking inside every cited URL, authority scores, and risk flags |
| Content Writer & Optimizer | Research-backed article production and gap-based page optimization, with inline comments and a real editorial process |
| Agents | Automated workflows over your data — recurring analyses, comparisons, and reports |
| MCP server | Your data inside Claude, ChatGPT, and any MCP-compatible assistant — see the MCP section |
Where the data comes from
Everything you see is measured, not estimated: real prompt runs against the live engines, real answers parsed for mentions and sentiment, real citations resolved to their source pages. When a number needs justification, the platform can show the exact answer excerpt behind it.
Start with the prompts guide — a focused portfolio of 5–10 prompts teaches you more in week one than fifty scattered ones.