Agent Mode
Agent is the AI Assistant's flagship research mode and the new default. Instead of answering in a single pass, Agent runs a structured investigation — it plans the research, reads your matter documents, calls case law and web search when external authority is needed, and synthesizes a legal memorandum with inline citations that link back to the source.
Agent is selected by default in the AI Assistant. Research and Deep Research remain available as legacy options in the mode selector.
What you see
Show Thinking panel opens
When you submit a question in Agent mode, the Show Thinking panel opens so you can watch the Coordinator plan the investigation and surface the Key Issues it intends to pursue.
Leads run in parallel
Each lead is a line of inquiry — a case to validate, a document section to read, a statute to retrieve. You see leads start, finish, and roll up into facts as they complete.
Memo is written
A structured legal memorandum is assembled with inline superscript citations that link directly to the source.
Sources & Citations panel
Every retrieved source appears in the Sources & Citations panel — grouped by Case Law and Web, with court metadata, a search box, and a Copy All button.
When to use Agent
- Analyzing matter documents ("summarize the contract risks in this MSA")
- Validating case law ("pull and verify Trevino v. State, Formosa, and Kroger")
- Any question that combines matter context with external legal authority
- Any time you need a cite-ready answer you could hand to a partner
What's different from Research and Deep Research
| Capability | Agent | Research (Legacy) | Deep Research (Legacy) |
|---|---|---|---|
| Transparent investigation (Show Thinking) | Yes — live leads and key issues | Limited | Limited |
| Inline citations | Linked to CourtListener opinions and page-level PDFs | Source links only | Source links only |
| Matter-aware by default | Yes | No | No |
| Output shape | Structured legal memorandum with inline citations | Chat response with source links | Long-form narrative response |
| Speed | Medium — scales with query complexity | Fast | Slow |
Three things Agent does that the legacy modes did not:
- Transparent investigation. Watch the AI plan and execute the research in real time — not just the final answer.
- Inline citations tied to source. Case law citations link to CourtListener opinion pages. Document citations open the source PDF at the cited page.
- Matter-aware by default. Agent knows which matter you're in and prioritizes its documents.
Example prompts
- "Review the attached Purchase Agreement for enforceability risks under Texas law and cite the controlling cases."
- "Extract all case citations from this matter's documents and validate them."
- "Draft a timeline of the payment disputes in this matter with supporting document citations."
What Agent is not
- Not a courtroom-ready document. Always review citations and holdings before filing.
- Not for quick conversational Q&A where you don't need rigor — Research is faster for a simple lookup.