Frame the research question
Write the decision the research should support and the audience for the output. Split the broad topic into a small set of questions that can be answered with evidence.
- What must be true for the decision to move forward?
- Which terms need precise definitions?
- What date range, geography or market applies?
- What evidence would change the current assumption?
Discover and record sources
- Use search or an AI assistant to identify candidate sources and terminology.
- Prioritise primary documentation, original research and directly accountable organisations.
- Record the title, publisher, URL, publication or update date and the question each source may answer.
- Add at least one independent source for claims where incentives or interpretation matter.
Verify what each source supports
- Read the full relevant section, not just a search snippet.
- Capture the exact quotation beside your paraphrase.
- Check sample, method and date for research findings.
- Mark conflicts between sources instead of silently choosing the most convenient answer.
- Label anything not checked as unverified.
Synthesize without flattening uncertainty
Group findings by the original research questions. Separate established facts, interpretation and open questions. Ask an AI tool to test whether your synthesis omits a source or overstates certainty, but compare its critique with your source log.
Review the output before sharing
- Every important claim has a working source link.
- Quotations, dates, names and calculations match the originals.
- Recommendations follow from the evidence rather than a generic best-practice list.
- Limitations and unresolved questions remain visible.
- Sensitive material and client information have been handled appropriately.
Official and reference links