Use this for discovery-heavy tasks where you need context and summary output.
Workflow
- Seed search queries.
- Run broad discovery.
- Deepen only the top matches.
- Synthesize into briefs.
Keep discovery broad, then spend deeper research credits only on promising rows.
Claude Code Prompt
Use research-queries.csv and build a two-pass Deepline research workflow:
1) broad discovery for each query
2) deeper research only for the top matches
3) synthesize a short brief per row
Run rows 0-3 first and write output to research-queries-output.csv.
Return:
- the action chain you used
- evidence links used for each brief
- a cost-conscious scaling plan for rows 4+
Why This Layout Is Efficient
- Async depth is isolated to high-priority hits.
- You avoid overpaying for unnecessary deep research.
- You can route the output directly into review workflows.
