What AI Radar measures
Radar asks the questions you want to be cited for, sending them to five AI engines every week, then reads back which sources each engine actually cited. It measures real citations. Each week opens with a plain-language Overview in three bands: the read (is AI recommending this brand), what moved, and this week's work. The Search Console style detail sits under it in Slice the numbers, and the work itself lives on Moves and the Action plan.
AI engines regularly recommend Fernway Coffee.
First source cited, on the same 128 buyer answers this week. Measured on your questions, not an industry average.
- The trend chart. The citation rate over time, with every engine drawn as its own line, the read week marked, and a marker on each week you shipped work. The live Overview flips the same chart to month, quarter, or a custom range.
- The verdict. A plain sentence: is AI recommending this brand, with the cited rate and its honest range.
- Methodology chip. Single-run reading or 3-pass confidence band, always visible.
How to use it
Add your domain.
Confirm your market and awareness stage.
Let the first weekly scan run.
Open Overview to see your citations, rate, and trend.
Work through Moves, one lost answer at a time.
Read the email digest each week.
How to read it
The headline number is your citation rate. The trend tells you whether it is rising or falling. Anything flat or empty is where engines answer without you: Moves turns each of those into a checklist, and the Action plan holds the account-level work.
Good to know
AI Radar measures real citations. The Scanner's AI Citability Index predicts readiness from on-page signals instead. Use the Scanner to prepare a page, and Radar to track whether it worked.
Try Radar with a free check, then track a domain every week on a paid plan. See Radar pricing.
If you know Search Console, you know Slice the numbers
Slice the numbers, the drill-down at the foot of the Overview, borrows the Search Console layout on purpose: the same answers cut by page, engine, buyer stage and appearance, over any date range. If you know Google Search Console, its four headline numbers map one to one.
- Citations
- Like Clicks. How often an engine cited you.
- AI answers checked
- Like Impressions. How many AI answers we looked at.
- Citation rate
- Like CTR. The share of answers that cited you.
- Avg cite position
- Like Position. Where you tend to sit among the cited sources.
How to read it
Read Slice the numbers the way you read Search Console. A rising citation rate is the goal. The plain-language Overview summarises the same data without the vocabulary.
The free AI Radar check
The free check is a one-time taster. It asks eight sample questions across all five engines and emails you the result. You do not need an account to try it.
How to use it
Go to the AI Radar page and enter your domain and email.
Submit the check.
Wait about three minutes while it runs.
Open your result from the email.
How to read it
The free check shows your citation score, a per-engine grid, and a competitor leaderboard. It is a snapshot. Paid tracking runs every week so you can watch the trend.
The free check needs no account. To track a domain every week, see Radar pricing.
Setting up tracking
Setting up tracking tells Radar who your buyers are and where they search, so it can build the right set of questions to track. For the full walkthrough with real screenshots of every screen, see the step-by-step setup guide.
- Market / region
- The country and language Radar asks from, so results match where your buyers actually search.
- Awareness stage
- Whether your buyers are just discovering the category, comparing options, or ready to choose. It shapes the kind of questions Radar tracks.
How to use it
Add the domain you want to track.
Pick your market and language.
Pick your awareness stage.
Review the questions Radar suggests, then edit or add your own.
Weekly tracking is part of the SMB and Agency plans. See Radar pricing.
The dashboard, screen by screen
Home is your client list. Each client opens into its own workspace, and every screen in it does one step of the same loop: see where you stand, do the work, confirm you did it, and learn whether it worked.
- Clients
- The home screen: one card per tracked domain, sorted so the clients that moved this week float to the top.
- Overview
- Start here. Three bands at whatever period you pick (week, month, quarter, a custom range): the read (the trend chart with every engine as a line and the work you shipped marked on it, then the verdict and a head-to-head against your strongest rival), what moved (every question that climbed or slipped, and who holds the answers you are missing), and this week's work (what is waiting on Moves and the Action plan). Slice the numbers sits at the foot.
- Competitors & sources
- Who holds the answer slots, over any date range: every rival domain, its share of the slots, and the exact thread or page each placement sits on.
- Moves
- A checklist per lost AI answer. Each move names who holds the slot, what the engine said and the steps to take it back. Tabs: To do (answers you are not in first, then answers you are in but below first), Pending (waiting on a reply to an ask, or on the next scans), Complete (decided by the scan), Past asks (declined or unanswered, suggested again later) and Dismissed.
- Your site
- The map Radar keeps of your site: what each page is for, and which page should answer each tracked question, including the ones that need a new page or a new section.
- Queries
- The buyer queries you track: add, edit, prioritize, and manage the bank. How each query performs (and exactly why it is not cited) lives on the Overview.
- Action plan
- The ranked account-level work this cycle, and the foundations under it, in three tabs. To do holds this cycle's actions, each anchored to the data point that triggered it, with anything that came back after you marked it done at the top. Foundations holds the groundwork phases. Done keeps what you finished, with its date.
- Client reports
- The archive: preview, curate, and publish the white-label weekly and monthly documents to a durable client link.
- Client settings
- This client's scans (live status, rescan, schedule), profile, market, awareness stage, GA4 connection, pause, and archive.
- App settings
- Account-level: white-label branding and model versions. Reads Agency settings on white-label plans.
How to work a move
Open Moves and start with the first row under To do.
Read who holds the slot and what the engine said, then tick each step as you do it.
Confirm it is complete once every step is live. For an ask (a listing or a directory), log the reply instead.
The next two weekly scans decide it: the move lands under Complete as won, or as cited on some runs.
On the Overview's query table, every tracked query is marked with one of three statuses:
- Cited
- An engine answered the question and pointed to your site as a source.
- Mentioned
- An engine named you in the answer but did not cite your site as a source.
- Absent
- An engine answered without mentioning or citing you.
How to read it
Start on the Overview for each client, work through Moves and the Action plan, then publish the Client report. You can see a full sample client report with fictional data.
The five engines
Radar checks the same five AI engines on every plan, with no per-engine add-ons.
- ChatGPT
- Perplexity
- Claude
- Gemini
- Google AI Mode
How to read it
Coverage differs by engine. It is normal to be cited often on one engine and rarely on another. The engine lines on the Overview chart, and the Engines cut in Slice the numbers, show the split so you know where to focus.
Weekly scans and email digests
Radar re-asks your questions every week and emails you a digest when the citation picture changes. It is weekly, not real-time, by design. AI citations move slowly, so a weekly read is the honest signal.
The trend across weeks is the signal, not any single week.
How to read it
Watch the trend across weeks, not a single week. The digest highlights what moved. For why weekly beats daily, see weekly vs daily AI citation tracking.
Single-pass vs the 3-pass confidence band
How often Radar asks each question depends on your plan.
- Single-passPaid
- Each question is asked once a week. A clear, honest read for one domain.
- 3-pass confidence bandAgency
- Each question is asked three times a week and shown as a range, so a one-off answer does not look like a trend.
How to read it
Read a band as a range. If the band is wide, the engines were inconsistent that week. For the reasoning, see the 3-pass confidence band.
Competitors and managing prompts
Two screens shape what you track. Competitors & sources shows who beats you, over any date range, and Queries is where you manage the questions.
- Share of voice
- Your portion of the cited slots versus competing domains, for your tracked questions.
- Google AI slots column
- On the Queries tab: how many sources Google's AI answer cites for that query in your market. Those are the slots you are competing for, so it describes the query's demand, not your result. How you are performing on each query is on the Overview.
How to manage prompts
Open the Queries tab.
Add a question your buyers actually ask.
Drag your most important questions to the top.
The top questions are the ones scanned every week.
How to read it
Use Competitors & sources to find a question a rival owns, make sure that question is in your tracked set, and Moves picks it up on the next scan.
Multiple domains and the white-label client report
Agencies track multiple client domains from one home screen, and each client gets a weekly white-label report on one durable link: your logo and name on it, past weeks archived inside it, and each published week's numbers frozen. See the sample client report.
How to use it
On Agency, add a client domain.
Set its market and awareness stage.
Apply your branding in Settings.
Open the client's Client report screen, review the preview, and publish.
Share the link once. New weeks publish to the same link automatically.
Multiple domains and white-label reports are part of the Agency plan. See Radar pricing.
Key takeaway
Add your domain, confirm your market, and let the first weekly scan run. Then read This Week for each client, act on the Action plan, and publish the client report.