At the end of April, I had the chance to attend AI x KM: The Next Frontier in Legal Knowledge Management, an event hosted by Inside Practice in New York. I came away knowing what I had already suspected - legal is leading the AI race.
Few functions are more poised for transformation in the legal world than knowledge management. And at AI x KM, that transformation was part of every presentation and discussion.
Every session offered a couple of gold nuggets. Here are the key takeaways that stuck with me, and I’d love to hear yours in the comments.
Legal is leading
We talk a lot about how risk-averse the legal industry can be. But when it comes to applied GenAI, law firms are ahead of the curve. That was confirmed by Bart Czernicki, Principal Global Black Belt, Artificial Intelligence at Microsoft. He also shared OpenAI's "Five Levels" framework and Microsoft’s Work Trend Index (see useful resources below).
AI & KM: inseparable partners
If there was a core message of the day, it was this: AI and KM go hand in hand.
KM teams are no longer just stewards of precedent documents or intranet pages. They’re leading AI strategy, surfacing insights, improving lawyer workflows, and becoming strategic enablers of transformation. The line between Innovation, KM, and IT is blurring fast.
Change = Risk (and that's okay)
One line I really appreciated came from Meredith Williams-Range, Chief Legal Operations Officer at Gibson Dunn: "AI = change and change = risk."
It’s simple, but powerful. Change brings risk, but also opportunity. For law firms, the challenge is how to balance that risk while still moving forward.
AI & client expectations
There was some interesting discussion about Outside Counsel Guidelines and GenAI use. Some clients (especially in tech) encourage it. Others have restrictions.
The answer? Dialogue. Law firms must help clients explore where AI adds value and how that value is safely delivered. It's not just a tech decision; it's a relationship one.
The Six Spheres of AI fear
I appreciate methods that simplify complex concepts. The six spheres was mentioned but not specifically what they are. So after a little research, I found these: Job Displacement, Economic Disruptions, Ethical Concerns, Societal Concerns, Existential Risks, and Long-term Risks.
These spheres encapsulate the anxiety and complexity experienced by firms and humanity. It is imperative that we develop improved frameworks for navigating these challenges.
Symbolic Agents and Decision Intelligence Frameworks
I especially enjoyed the discussion with Dr. Megan Ma, Executive Director of Stanford Legal Innovation through Frontier Technology Lab (liftlab) at Stanford Law School, Ilona Logvinova, Director of Practice Innovation at Cleary Gottlieb Steen & Hamilton LLP and Damin Riehl from V-Lex. I learnt a couple of new things.
I had not heard of symbolic agents before. They represent knowledge-based expert systems, using predefined rules to solve complex problems. The user sets the goal, humans prompt common tasks, and LLMs execute those tasks. This is different to LLM Agents where the user sets the goal and then the LLMs both choose and execute tasks.
Turns out when doing further research there are also Reactive Agents, Reinforcement Learning-based Agents, Empathetic Agents, Philosophical Agents etc.
Agents, role play & machine colleagues
We’re entering the era of AI agents as machine colleagues. These tools won’t replace lawyers — they’ll assist, draft, review, and even ask questions before we think to.
One fascinating concept discussed: training AI agents via interview. With just 2-5 hours of dialogue, teams are capturing:
- Memory (war stories)
- Muscle memory (workflow triggers)
This yields 85% of what makes a lawyer unique and provides the kind of context that AI needs to be useful.
Experiential learning
Caitlin Vaughn and Anne Stemlar from Goodwin Procter presented a 9-week experiential learning program for new associates that had great outcomes for the firm. Naturally, it generated lots of questions from the audience.
As AI reshapes workflows, this kind of onboarding will become even more critical and training has to go beyond how to "use a tool".
Modern KM
One of the clearest signals from the event was that KM is no longer about curated documents. Modern KM is about:
- Leveraging technology to perform traditional KM tasks more efficiently and at scale.
- Well-organized data that can effectively feed and train AI.
- Power pricing, client insights, and operational intelligence.
As one speaker said: KM isn't about documents. It’s about enabling better decisions.
Embedding, not just evaluating
And the final session was of course the best session (my unbiased view 😃) with:
- Sarah Hirebet (Andrews) from Stradley Ronon
- Sarah Pullin from Baker McKenzie
- Carolyn Austin from K&L Gates
- Moderated by Gabriel Karawani (Atlas by ClearPeople)
Sarah delivered a much-needed reminder to be brave enough to stop evaluating new tools and embed what you've already introduced. It’s tempting to chase the next GenAI launch, but law firms (and lawyers) can only handle so much change. In fact, 3 major changes per year is the threshold most teams can absorb.
Sarah Pullin shared how alignment across KM and IT has been critical to their success — not just in deciding what to pilot, but in moving strategically from POC to scale.
Carolyn Austin noted that law firm partners are actively approaching KM teams for AI input — a sign of how central KM is becoming to AI strategy.
Final thoughts
AI x KM was a brilliant reminder of how far legal has come — and how much farther we can go if we get KM right.
We're not just watching the AI future unfold. In legal, we're helping shape it.
Useful Resources
