Start with one buying situation.
Choose one market, one language and one priority audience. Write down the decision that audience needs to make: choosing a sofa for a small apartment, comparing managed IT providers or finding cookware for an induction kitchen. A focused buying situation makes the findings easier to interpret.
Use sales conversations, support questions and your existing customer research to identify needs. Label analyst-created questions as research questions; do not present them as measured customer demand.
Separate discovery from branded evaluation.
Discovery questions describe a need without naming your brand. Branded questions ask about a named business or compare specific options. Keep these groups separate so recognition prompted by your name does not inflate your view of unprompted discovery.
- Discovery: “Which sofa brands should I consider for a small apartment?”
- Branded evaluation: “What should I check before buying an Aster Home sofa?”
- Comparison: “How do Aster Home and Form & Field compare for a narrow living room?”
- Purchase checks: “What are the delivery and return terms for Aster Home?”
Give every question a job.
Tag each question with its buying stage, audience, product group and research objective. Decide what you will record before collecting answers: a brand mention, shortlist inclusion, an explicit recommendation, a factual claim or a cited source.
Keep questions natural and specific. Do not add your preferred answer or a claim you want the system to repeat. Review near-duplicates so one wording pattern does not dominate the set.
Record the collection conditions.
Keep the exact question text alongside the AI experience, collection date, market and language. Record available model or version information, search settings, session context and sampling pass. Save the full answer and any visible source links.
If you repeat the research, preserve a stable baseline set. Record new questions and changed conditions separately. A difference in answers needs context before it becomes a trend.
Turn the set into a research brief.
Name the decisions the findings must support. Agree which competitors matter and which facts need verification. Set a review date and decide who will interpret the evidence.
Use the question-set template to document the scope and collection fields before you begin. For a specialist assessment, Lens defines the question set and delivery scope with your team.