Deep Research
Produce memo-quality research reports with multi-source iterative AI research.
Deep Research is an advanced research mode that produces structured, memo-quality reports by searching across multiple sources and iteratively deepening its analysis.
Deep Research is off by default. Toggle it on in the AI Assistant panel before submitting your query. Responses take longer than standard search but are significantly more comprehensive.
How it works
Multi-source search
The AI searches across four source categories in parallel:
Internal documents
Files uploaded to the current Matter (contracts, memos, exhibits).
Case law
U.S. case law via CourtListener's 50M+ citation database.
Web
Public legal resources, law review articles, and government sites via Tavily.
Statutes
Federal and state statutory databases.
Iterative deepening
After the initial search, the AI evaluates what it found and identifies knowledge gaps -- areas where the results are thin or conflicting. It then runs additional targeted queries to fill those gaps automatically.
This cycle repeats until the AI determines it has sufficient coverage to answer the question comprehensively.
Fact cache
Deep Research builds a fact cache as you ask follow-up questions within the same session. Prior findings are retained so:
Follow-up queries start from a richer knowledge base.
The AI avoids redundant searches for information it already retrieved.
You can refine or narrow your research incrementally without starting over.
When to use Deep Research
Quick case lookup
Recommended
Overkill
Single-issue legal question
Good
Better
Multi-element analysis (e.g., "elements of fraud in Texas")
Limited
Recommended
Memo or brief preparation
Insufficient
Recommended
Cross-jurisdictional comparison
Limited
Recommended
Deep Research responses take longer to generate due to iterative search cycles. For quick lookups, use standard Case Law Search instead.
Last updated