Insights
Field notes on legal intelligence: the legal ontology, knowledge graphs, governed agents, and the craft of making a firm's knowledge compound.
Subscribe via RSSHow we use AI agents at Reframe
How we build with AI agents at Reframe: the context they start from, cited to the source, the permissions and walls that bind them, and the changes a person approves. What it means for a law firm.
The billable-hour problem with legal AI, and what firms are doing about it
AI that makes lawyers faster reduces hourly revenue, and firms are paying for it. How leading firms are rethinking pricing, AI cost models, and what the firm keeps.
Your lawyers are building software now. How IT can make it safe.
Lawyers are building their own tools with AI. How a CIO can find what exists, set tiers and owners, and support the tools worth keeping.
When a law firm builds AI products for its clients
Clients are asking firms for AI tools they can use directly. The four decisions to make before the first client signs on.
The AI committee most law firms do not have
Most firms route AI through a technology committee built to choose tools. What an AI committee should decide, who belongs on it, and how to stand one up in 90 days.
Why law firm counsel are pulling AI back inside the firm
Counsel at large firms are losing patience with AI vendors that learn from firm data, even anonymized. What is driving the shift, and six questions for every vendor.
Privilege after Heppner: what a law firm CIO should be ready to answer
Four decisions this year turned AI privilege into an infrastructure question. What they say, and what the GC will ask IT.
Before AI, a law firm has to know which matter every document belongs to
Walls, access rules and retention all hang on client and matter numbers. Why attribution comes first, and how to enforce it at intake.
The people a law firm needs to run AI, and where IT fits
Firms have budget for AI roles and still can't fill them. The four roles that matter, and how IT moves from cost center to technology first.
How to keep partners behind an AI program long enough for it to pay off
Programs end when partners stop believing in them. Dated targets, early client use and plain-language reporting keep them alive.
The firm's knowledge is the asset. Who holds it matters.
What your lawyers teach AI every day is the firm's judgment in writing. Five contract terms that decide who keeps it.
What belongs in a law firm's 2027 AI strategy
Most 2027 innovation strategies are a list of tools. Here are the decisions a law firm's AI strategy should record, and seven questions to test a draft.
Why most law firms don't need a data lake before AI
Many firms are starting a data lake as the first step of their AI strategy. For most law firms, a living ontology is the better starting point, and it puts agents to work now.
How IT and the business can build a law firm's agentic strategy together
CIOs want an agentic strategy as much as anyone. How IT and the business build it together, and why capturing what lawyers know needs a legal ontology.
AI for law firms: the complete guide
What legal AI can do now, where it breaks, and why grounding it in a legal ontology is the difference between a demo and a system.
How to evaluate a legal AI vendor
The buying criteria that actually matter: grounding, security, deployment, governance, and whether it is built on a legal ontology.
When a partner leaves, the firm should keep what they knew
Retirement takes decades of a partner's judgment with it. What actually keeps the reasoning inside the firm.
Every closed matter should make the next one cheaper
Why matter exhaust should compound like capital, and what structuring it actually takes.
An AI adoption playbook for firms that bill by the hour
Six moves that survive billable-hour economics, risk committees, and partner skepticism.
Agents can do legal work. Governance decides whether they should.
Scoped permissions, approval gates, and audit trails that make agents deployable at a firm.
The cure for hallucinated citations is grounding, not hope
Grounded retrieval and source-linked answers end fabricated authority. Better models alone do not.
Closing conditions do not slip because lawyers are careless
A live obligations register that keeps conditions, consents, and covenants on schedule.
An AI-assisted due diligence checklist for M&A
Scoping the data room, review tiers, red flags, and verification discipline for real deals.
Drafting with AI where attorneys actually work: in Word
What attorney-grade drafting requires when the AI lives inside the document itself.
Knowledge management is finally becoming infrastructure
When AI reads every matter, knowledge management stops taxing your busiest people.
Your DMS is the system of record. AI should treat it that way.
Permission-aware ingestion and query-time enforcement for iManage and NetDocuments.
AI contract review for law firms
Review that runs in your own cloud changes accuracy, privilege posture, and the client consent conversation.
The SOC 2 questions legal AI vendors hope you skip
Type II scope, subprocessors, retention, and the answers that actually matter.
Where legal AI runs is the security question that decides the rest
Keys, networks, and query-time control: what a dedicated, isolated enclave in your own cloud really means in practice.
Private legal AI vs public chatbots: what actually differs
Privilege, retention, grounding, and audit: the four gaps that decide the question.
RAG or knowledge graph? For legal work, the answer is both.
Passage retrieval finds text. Graphs understand relationships. Serious systems use both.
A legal ontology, explained without the jargon
The shared vocabulary of entities and relationships that lets software reason about legal work.
What a legal knowledge graph actually does for a law firm
How matters, documents, and precedent become a queryable institutional memory.