Answer Radar
Choosing and Approving Buying Questions
Buying questions, called prompts in the tools, are what Answer Radar asks the engines. Choosing them is the single most important decision in the product, because a score is only ever a measurement of the questions behind it. The Questions page gives you three sources of questions and four ways to shape them, and it audits only what you approve.
Quick Answer
On Questions, generate a proposed set, approve the questions worth measuring, remove the rest, and add your own. Classify type and funnel stage and tag topics to organise them. Approval is separate from labelling, and only approved questions are audited.
- Generate questions proposes commercial-intent buying questions into a draft set.
- Add your own buying question puts a question you wrote into the same set.
- A set becomes active as soon as it holds one approved question.
- The Run audit button refuses a project with no approved questions.
On this page
Three sources of questions
Proposed: press Generate questions and Answer Radar writes commercial-intent buying questions for your project into a draft prompt set, using your domain, category, competitors and buyer segment. Press Regenerate for another set. How that proposal works is its own article.
Your own: 'Add your own buying question' puts the exact wording you typed into the set. Use it for the questions you already know buyers ask, especially ones taken from Search Console once it is connected.
From an MCP client: answer_prompts_update with add does the same from Claude or Cursor, and answer_prompts_propose generates a draft set the same way the button does.
Approve, remove, reprioritize
Only approved prompts are audited. Approve the questions worth spending credits on; remove the ones that are not, which is a soft removal; and reprioritize to control order. The page tells you when nothing is approved: 'No approved questions yet. Generate a set, then approve the ones worth auditing.'
Prompt sets have a status of draft, active or archived. The browser and the MCP tools both promote a set to active once it holds an approved prompt, and never demote it when approvals are removed, so the two surfaces leave indistinguishable state.
Classify and tag, without approving
Classify labels a question's shape and funnel stage. Type is one of best_of, comparison, alternatives, pricing, how_to or review. Intent is free text such as bofu. One value applies to every question you select, and an omitted field is left alone. Nothing generates a type: you read the prompts, decide the labels and write them back.
Topics are your own grouping. Every topic you add goes onto every question you select, and remove takes it off. A topic exists only because prompts carry it, so there is nothing to create or delete; to rename one, list its prompts and send them back with the new name added and the old one removed. Metrics and trends can be filtered by topic.
Classifying and tagging do not approve anything. Labelling and deciding what gets measured are deliberately separate acts.
What makes a good set
Questions a buyer would ask before choosing, phrased the way they would type them: best tools for a job, alternatives to a named competitor, comparisons, pricing, how to do the task. Avoid questions that only your brand name could answer; a question that names you measures recall, not recommendation.
Keep the set small enough to afford daily and large enough to mean something. The configured maximum is 50 questions per run. If Search Console is connected, the real queries buyers used to reach your site are the best sanity check on a proposed set.
- Cover each type at least once: best_of, comparison, alternatives, pricing, how_to, review.
- Tag by product line or persona so metrics can be filtered by topic.
- Approve deliberately; every approved question costs one credit per engine per sample.
FAQ
Can I edit a question's wording after approving it?
Remove it and add the corrected wording as a new question. A prompt's text is snapshotted into every run that measured it, so history stays attached to exactly what was asked.
Does removing a question delete its history?
No. Removal is soft, and every sample already measured for it stays in the runs that recorded it. Evidence is append-only.
How many questions should I approve first?
Enough to cover your main buying questions across the six types, and few enough that the first audit's estimate is comfortable. Ten to twenty approved questions on four engines is a common first run.
Related help pages
Answer Radar
How Answer Radar Proposes Buying Questions
What Generate questions does, what it draws on, why proposals land in a draft set you approve, and how to ground a set in real Search Console queries.
Answer Radar
Segments and Topics: How Results Are Grouped
Why every metric is reported per segment and never averaged across engines, what the five axes are, what the instrument token is, and how topics group prompts.
Answer Radar
Connecting Google Search Console
The service-account flow: add Linkeddit's address as a user on your Search Console property, check for properties, pick one. No OAuth, read-only, and what.
Answer Radar
The Audit Run and What It Costs in Credits
What Run audit does: the estimate, the reservation, the fan-out per question per engine per sample, the run states, and why one credit is one measured sample.