Knowledge Management for Lawyers Working With AI

Knowledge management for lawyers is the practice of capturing and making reusable what a firm or attorney has learned: research on recurring legal questions, precedents from prior work, approved process steps, and handoff context for active matters. Done well, it reduces time spent re-researching the same ground and helps the next attorney on a matter understand what was decided and why. For lawyers using AI assistants, it also solves the blank-slate problem — each new session starts without memory of the last one.

A tool like fenbs keeps approved working rules and nonconfidential project context in one place that connected assistants read before they help — without storing client communications, case files or any information that carries confidentiality obligations.

What knowledge is reusable in a law practice?

Not everything a lawyer knows is reusable, but quite a lot is. The checklist for a residential closing, the sequence of steps for incorporating a Delaware LLC, the research on whether a particular clause is enforceable in the jurisdiction you practice most — these can be captured once and used again.

• Practice area research: answers to questions that come up repeatedly, with the sources that back them.
• Process checklists: the approved steps for a recurring matter type, reviewed and updated when the law changes.
• Approved clause language: standard positions on common negotiation points, with notes on when to deviate.
• Matter context: the current status of an active file, decisions made, tasks outstanding and the next step — useful for picking up a matter after an interruption.
• Lessons from past matters: what worked, what was missed, what you would do differently on the next similar matter.

Client confidential information — communications, strategy, personal facts about clients — does not belong in a shared knowledge record. It belongs in your firm's matter management system with appropriate access controls. The boundary matters both for confidentiality reasons and because AI tools should not be handed data they do not need to do the job.

What do ABA ethics rules say about AI and client data?

The ABA Model Rules of Professional Conduct require lawyers to protect client confidences (Rule 1.6) and to supervise the tools and people working on their matters (Rules 5.1 and 5.3). The ABA's Formal Opinion 512, issued in 2024, addresses generative AI directly and requires attorneys to understand the tool they are using and to take steps to protect client information.

State bars have issued their own guidance. Check your jurisdiction's bar association for current opinions on AI use before deploying any AI tool in client work. Terms vary: some bars require disclosure to clients; others focus on competence and supervision. Do not rely on a general summary of the ABA Model Rules as a substitute for your jurisdiction's current guidance.

The practical implication is a clear separation: the AI tool you use for drafting, research or summarization should work from procedural and research context, not from client files. Keep those two categories separate from the start of a matter.

How do you organize reusable legal research?

Reusable research has a short shelf life if it is not maintained. A memo on whether a non-compete clause is enforceable in California was accurate when it was written; it may not be after the next state court decision. Build maintenance into the process: when you update research, note the date, what changed and whether any connected checklists or clause positions need updating.

Organize by practice area and question, not by matter. A matter-based filing system is useful for retrieving work done on a specific client; it is not useful for retrieving the answer to a recurring legal question. The two organizations serve different purposes and can coexist.

Link to primary sources. When you save research, include the case citation, the statute and the relevant regulatory guidance URL. An AI assistant working from a research note with primary source links can check the current status of a rule or flag that a citation needs verification. One working from a summary without sources cannot.

How do you handle matter handoffs with AI?

A matter handoff between attorneys is a knowledge transfer problem. The incoming attorney needs to understand the current status, what has been decided, what is still open and what the strategy is. Without a written record, that transfer depends on a conversation that may not cover everything and a review of files that does not explain the reasoning.

A brief where-we-left-off record solves much of this. After each working session, note what was done, what decisions were made, what tasks are still outstanding and what the next step is. When another attorney — or the same attorney after a gap — opens the matter, they have a current summary without having to reconstruct it from emails and documents.

For AI-assisted work, this record also solves the session-memory problem. Ask your AI assistant to save progress before closing a session: what was drafted, what was checked, what it found and what remains. The next session's assistant picks up from that note. Related reading: knowledge management strategy covers organization principles that apply across practice areas.

What does a good knowledge system for a law firm look like?

A small firm or solo practice needs a lightweight, maintainable system. A matter management tool holds client files. A research library holds reusable memos and checklists, organized by practice area. A project record holds active matter status and the decisions being tracked. Those three components cover most of what a legal knowledge system needs to do.

Larger firms may also maintain precedent databases, know-your-client records and firm-wide process documentation. Those require governance — someone responsible for keeping them current — and often involve dedicated legal knowledge management software. For coverage of the tools landscape, see knowledge management tools.

How to do this with fenbs

1. Create a free account. Start with a personal kanban board for your own matters, or use a team board if you share work with colleagues.
2. Open a project for a practice area or an active matter. Keep the project description focused on the nonconfidential context: the type of work, the process steps and the decisions the team has agreed on.
3. Write your approved process steps and standard positions in Decisions and rules. These appear first when a connected assistant reads the project context.
4. Add tasks for each active workstream: the next brief due, the research question under review, the document awaiting signature. Move them through the board as work progresses.
5. Before closing a session, save a Where we left off note: what was done, what was checked, what is still open.
6. After a matter closes or a research project concludes, write a Lessons learned note: what you would do differently, what the research confirmed or overturned.
7. Connect your AI assistant from Integrations and ask it to read the project context before it helps with any work on the matter.