Getting Started With Answer Radar

Step 1: check your plan

Answer Radar is available on the Compete plan and above. On other plans the pages render an upgrade notice reading 'Answer Radar isn't included in your plan. It's available on the Compete plan.' and the Answer Radar API routes return an upgrade response rather than data. Trials do not include it.

If you connect through MCP instead of the web app, the same entitlement applies: the answer_* tools check it before doing anything.

Step 2: create a project

Go to /answer-radar and choose a new project. The form asks for your domain, brand name, competitors, buyer segment, category, geography and language. Only the domain is required; if you leave the brand name blank it falls back to the domain. Everything else shapes how questions are proposed and how measurements are segmented, so fill in what you know.

Creation is idempotent. If the request is retried after a partial failure, the app finishes the half-built project instead of creating a second one. The full field-by-field guide is in the creating a project article.

Step 3: approve buying questions

After creating a project you land on Questions, because a project with nothing approved has nothing to measure. Press Generate questions to have Answer Radar propose commercial-intent buying questions into a draft set, approve the ones worth auditing, remove the rest, and add your own with 'Add your own buying question'.

Only approved questions are audited. Approving is a separate act from labelling: you can classify a question's type and funnel stage, or tag it with topics, without approving it. The Run audit button refuses a project with no approved questions and says so.

Step 4: run the first audit

From Overview or Questions, press Run audit. A dialog prepares an estimate: approved questions times enabled engines times samples per question, in credits, bounded by the per-run cap. Confirm and run. The audit reserves budget, fans out one request per question per engine per sample, and aggregates the results into gaps.

The run appears under History with a status: pending, running, partial, completed, failed, cancelled or budget_exceeded. A partial run is partial, not a set of zeros. Open a run to see every answer and every citation it recorded.

Step 5: read the results

Overview shows the latest audit's insights, the gaps browser and the audit history. A gap is a buying question where an engine recommended somebody else, or nobody. Each gap holds what the engines actually said, the sources they cited, hypotheses for why a competitor won, and recommended actions, kept strictly apart so a claim is never confused with a guess.

Actions is the queue of cards the audit justified. Fixes is where drafts built from the evidence are reviewed and approved. The metrics behind all of it are defined on the methodology page, and every figure carries the sample counts and versions that produced it.

• Overview: latest insights, gaps, share link, Search Console connection, run audit.
• Actions: content gap and competitor won cards with their evidence.
• Questions: propose, approve, classify and tag buying questions.
• Fixes: review, approve or send back drafted fixes.
• History: every run with status, sample count and credits.

FAQ