01 / Decide what the research needs to help you do
Define a question, audience, geography, and time period. Add the decision the report should inform and anything outside its scope. “Remote work” is broad; “What evidence should a small agency consider before changing its remote-work policy?” is easier to investigate.
Choose a research cutoff rather than using an old year embedded in a template. The updated prompt asks for that date explicitly.
02 / Agree on a research plan
Before producing a report, ask the assistant which subquestions it will investigate and which source types would answer them. If it lacks web access, provide sources and ask for a review of that material only.
Match the source to the claim. A product feature can be checked in official documentation; a claimed effect needs evidence suitable for that question. A source can be authoritative about its own offering without being an independent evaluation of it.
03 / Make an evidence trail
For the remote-work example, separate employee preferences, recruitment effects, coordination costs, and business outcomes. Those are different questions and may require different evidence.
Ask for a table containing each important claim, its source, the source date, the relevant passage or finding, and any limitation. A link that exists but does not support the claim is not a useful citation.
Get the exact prompt.
Make it your own.
You’ve seen how it works. Unlock the copyable prompt and the remaining steps to try it yourself.
- Copy the research workflow
- Read the uncertainty before the recommendation
New here? Create a free account