Use case · Suggested workflow
AI agent research workflow: build a source-linked shortlist
Content updated · Editorial guidance; not a hands-on product test.
An AI agent research workflow turns a decision into a source-linked shortlist: define requirements, collect options, verify disqualifiers, examine tradeoffs, and save the decision record. The steps below are a suggested method, not a completed product test.
1. Set the decision
Write one sentence explaining the choice you need to make. List hard requirements separately from preferences. For example, a tool must export CSV; a dark theme is only a preference.
2. Ask for a narrow shortlist
Request three to five options with a primary-source link for every material claim. Ask the agent to leave a cell blank when it cannot verify it.
3. Check the disqualifiers first
Visit the sources for price, availability, required features, and plan restrictions before reading general descriptions. Eliminate options that fail a hard requirement.
4. Ask for the strongest objection
For each remaining option, ask what would make it a poor fit. This helps expose differences hidden by similar feature lists.
5. Keep a decision record
Save the shortlist, the date, and the reason for your choice. Revisit the comparison when your requirements change. This is an editorial workflow, not a report of a completed product test.