Get named in AI answers. Not just measured.
Prospects ask ChatGPT, Claude, Perplexity and Gemini what to use. Radar measures whether you make the short list, and Operator drafts the work that gets you on it.
ChatGPT
Claude
Perplexity
Gemini
Google AI OverviewsHow does Radar measure whether AI assistants recommend me?
Radar keeps a set of tracked prompts, the questions your prospects ask, and each week puts every one to ChatGPT, Claude, Perplexity, Gemini and Google AI Overviews, one to ten runs per prompt. From the answers it records two separate rates per engine: how often you are named, and how often your site is cited. Each rate carries a 95% Wilson confidence interval, and a week-over-week change is only flagged significant when the two intervals do not overlap. Operator then acts on the gaps. It drafts posts for a blog on your own domain, which wait for your approval, and drafts listing requests to the pages engines already cite. Radar is in beta and is enabled account by account.
Every mention rate and citation rate is reported with a 95% Wilson interval, per engine. A movement is marked significant only when this week's interval clears last week's.
A third-party site becomes an off-site target only once it is cited in at least three distinct answers. Up to eight targets are recommended per run.
Puts the blog Radar publishes to on your own domain. Cloudflare issues the SSL certificate automatically once the record resolves.
5 engines, separately
ChatGPT, Claude, Perplexity, Gemini and Google AI Overviews, never blended into one score.
Honest intervals
Mention and citation rates with 95% confidence intervals, not a single spot check.
Drafts the fixes
Posts for your own domain and listing requests to cited pages, all approval-gated.
Watches competitors
Tracks share of voice so you see when a rival gains or loses ground.
A cycle you can watch, step by step.
Confidence intervals, not coin flips.
Ask an engine the same question twice and you can get two answers. Radar reports each rate with an interval and only calls a change significant when the intervals separate, so you act on movement rather than noise.
Named and cited are different wins.
Being mentioned in an answer and being the page it links to call for different work, so Radar tracks them separately, per engine, along with competitors' share of voice.
Findings become drafts you approve.
Operator drafts evidence-grounded posts for the blog on your domain and publisher-framed requests to the roundups and directories engines already cite. When a target site starts citing you, Radar marks it done.
The questions prospects ask before they switch, answered with evidence
Prospects put the same questions to ChatGPT and Perplexity every week. This post answers the three where your category comes up and you do not, each claim tied to a source engines already cite.
Built to run, not to sit in a dashboard.
Acts, not just reports.
Radar findings become Operator work on the same platform: drafted posts and drafted outreach to the sources engines cite.
Statistically honest.
Rates come with intervals, and a change is significant only when the intervals separate.
Content on your domain.
Posts publish to a blog on your own domain, so the authority accrues to you.
The details, answered.
What is answer engine optimization?
Answer engine optimization (AEO, also called generative engine optimization or GEO) is the work of getting your product named and cited when people ask an AI assistant for a recommendation. Radar measures where you stand and Operator drafts the work that moves it.
Which engines does Radar track?
ChatGPT, Claude, Perplexity, Gemini and Google AI Overviews, each tracked separately because they draw on different sources. Mentions and citations are tracked separately too.
Why do my results look different every time I check?
Because engines vary from run to run. Radar can run each prompt up to ten times a week and reports a confidence interval, so you see a range and only get told about a change when it is larger than the noise.
Do I have to write the content myself?
No. Operator drafts posts grounded in evidence for the blog on your domain. Nothing publishes without your approval, and you can edit any draft first.
How many prompts can I track?
10 on Starter, 50 on Growth and 200 on Scale. Radar is in beta and is switched on account by account while it is.
Does this replace SEO?
No, it builds on it. AI engines lean on search indexes and the pages they already rank, so crawlable pages, structured data and content worth citing serve both.