Best Practices
- 1Give it a specific research question, not an open-ended explore — scope it first
- 2Always specify what decision the research informs — it shapes the report format
- 3For API capability audits, ask it to verify claims against live docs, not memory
- 4Pearl_Research reports feed into Pearl_Marketing briefs and Pearl_Int integrations
- 5Ask for a confidence rating on each finding — it distinguishes verified from plausible
CLI Examples
# Competitive research act as pearl_research — what are the top 5 self-help audiobook platforms and their royalty models? # API capability audit pearl_research: what TTS voices does ElevenLabs support for ja-JP, and what are the rate limits? # Technology evaluation pearl_research: compare Metricool vs Buffer for 14-locale social scheduling — cost, API access, limits # Market analysis pearl_research: what is the addressable market for mindfulness books in zh-TW in 2026?
Do Not Use For
- Building anything — Pearl_Research produces reports, not code
- Replacing Pearl_Int for API wiring — research tells you what to wire, Pearl_Int wires it
- Treating research outputs as final without verifying against live sources