The AI agent lives in the right panel of every case. It searches your documents to answer questions, locate files, and run case management tasks. Every document-backed response includes AI citations so you can verify each answer against its source.
What the AI agent can do for document review
The AI agent handles a wide range of document review tasks. Instead of switching between screens and menus, you accomplish most of your work through natural language in a single conversational interface. Here are the agent's core capabilities:
- AI research and analysis - Ask questions about your documents and get detailed answers with source citations.
- Search and file discovery - Find specific files or information across the entire case using natural language.
- Redaction - Identify sensitive information and review proposed redactions before application.
- Bates numbering - Apply sequential Bates numbers to documents for production preparation.
- Tag management - Create, assign, and organize tags to categorize your documents.
- Custodian management - Create custodians and assign documents to them for tracking.
- Production exports - Generate production sets from tagged or filtered documents.
- Privilege marking - Designate documents as attorney-client privilege, work product, or both.
- Case memory - The agent retains findings and instructions across sessions.
- AI citations - Every response includes clickable references to source documents and page numbers.
For a full list of AI agent capabilities, see the AI agent capabilities page.
The AI agent experience
Type a question or command into the AI agent panel and press Enter. The agent responds with answers, AI citations, and completed actions. It streams responses in real time so you can follow along as results appear. When the agent references a document, click the citation to open that file in the document viewer at the exact page cited.
The AI agent maintains conversation context within each session. Ask follow-up questions that build on earlier answers without repeating background. For example, after asking about key contract terms, follow up with "Which of those terms were modified in the amendment?" The agent understands the context from your previous exchange.
For actions like tagging, redaction, or Bates numbering, the AI agent confirms what it will do before making changes. Redaction requires your explicit approval before application. Other operations like tagging and custodian assignment execute after the agent describes the planned action. You are responsible for reviewing all agent actions and correcting any errors.
AI agent case memory
The AI agent retains findings, conclusions, and instructions across sessions, so you do not need to repeat context each time you return to a case. Memory is shared with all case members. When one team member records a finding or provides instructions, the entire team can access that context in future conversations. Because memory is shared, team members should periodically review stored findings for accuracy. The agent may retain incorrect or outdated information.
Accessing the AI agent panel
The AI agent panel occupies the right side of the case page in Hintyr's three-panel layout. The left panel shows the document browser, the center panel displays the document viewer, and the right panel hosts the AI agent. This layout lets you browse files, read documents, and interact with the agent at the same time without leaving the case.
To start a conversation, type your question or command into the text input at the bottom of the AI agent panel and press Enter. The agent streams its response immediately. You can continue browsing documents or performing other tasks while the agent works. Response times vary depending on case size and query complexity.
To resize the panels, drag the dividers between them to give more or less space to the AI agent panel. Some users prefer a wider agent panel for intensive research, while others keep it narrow and focus on the document viewer.
Filtering AI agent results by tag
A tag filter dropdown at the top of the AI agent panel lets you scope the agent's work to specific tags. When you select one or more tags, the agent only searches and operates on files within those tags. Use this to focus on a particular document set, such as a production collection, a privileged set, or a group of files you already organized for review.
When no tags are selected, the AI agent searches across all files in the case. Selecting tags narrows the scope so results and actions only apply to files in those tags. You can select multiple tags, use the All button to select every tag, or use Clear to return to the full case scope.
Streaming AI agent responses
The AI agent streams responses in real time as it processes your request. You do not need to wait for the full answer before you start reading. As the agent works through your documents, progress indicators show search status, how many documents it reviewed, and what it found so far.
This streaming behavior gives you feedback and lets you assess whether the AI agent is on the right track. If the initial results are not what you expected, you can refine your question in a follow-up message. The agent uses your clarification as additional context for its next search.
Once the AI agent finishes, the complete response saves to your conversation history. Scroll back through previous messages at any time to review earlier findings or copy text from the agent's responses.