By Patrick DiDomenico, Founder and CEO, LegalTech Connect.
I have been using a phrase in conversations over the past year that I keep coming back to: ‘old problems, new tools.’ I started saying it partly out of frustration, watching new entrants to legal technology present longstanding knowledge management concepts as breakthroughs. But the recognition underneath the frustration is that the core of what we do in KM has not actually changed, even as the tools have gotten dramatically more powerful.
I spent this summer reading session proposals for the Knowledge Management and Innovation for Legal Conference (Oct. 21 to 22, NYC), which turns out to be a useful way to keep up with the realities of KM in a world reshaped by AI. The proposals did not look like the ones from previous years. But the problems underneath them were familiar.
What lawyers do for their clients has not changed much at all. They draft documents, negotiate deals, litigate disputes, advise on regulatory compliance, and help clients navigate risk. Those are the old problems, and they are the same problems lawyers were solving decades ago and will be solving decades from now.
The tools have changed dramatically. A lawyer who once pulled a precedent from a brief bank can now query an AI system that surfaces relevant clauses across the entire firm’s work product. A knowledge manager who once maintained expertise directories by sending annual surveys can now build an AI persona from a partner’s work product and structured interviews, so colleagues can engage with that judgment without the partner in the room. The capability is genuinely impressive, but the problem being solved is the same one KM has been working on for decades: capturing, surfacing, and transferring institutional knowledge.
Over the years we tried precedent collections, brief banks, expertise location systems, and intranets with an ‘ask the expert’ button that seemed like a solution, but never really caught on. Things improved and each generation of tools was better than the last, but the underlying problems stayed the same.
The hardest part of KM has never been the technology. It has always been getting busy lawyers to sit down and share what they know, and that participation problem has not gone away.
I have watched this cycle repeat across multiple technology generations. KM gets repositioned against IT, then against the library, then against innovation, now against legal engineering. Each redraw is usually the right call (or at least the best call at the time), but the firms that treat each boundary change as a fresh discovery tend to redraw it again in three years. The problem was never where KM sat in the org chart. It was always whether the firm was willing to do the unglamorous work of capturing knowledge before it walks out the door.
The same applies to the talent question. Firms keep looking for a unicorn who combines legal knowledge, business acumen, data literacy, and AI fluency in a single person, but that person mostly does not exist. An effective solution has been the same for many years: develop the people already sitting in your KM, library, and research teams; find the practicing attorneys who have a passion for helping solve these problems and recruit them into these positions. It is slow and creates friction when some make the leap and others do not. But so far, even in this advanced technological era, I have not found a shortcut.
I want to be clear about a distinction. The old problems I am describing are the problems lawyers solve for their clients, and those have not changed. But AI has created genuinely new problems that are different in kind, not just in degree.
Hallucinations and AI slop are real problems that did not exist before generative AI. A system that confidently invents case citations is not an old problem in new clothing. It is a new failure mode that requires new safeguards. The fact that lawyers have always verified citations does not make the hallucination problem trivial. It makes the verification step more critical and more frequent. And just to be clear, while hallucinations and AI slop are real, there is absolutely no excuse when lawyers fail to catch AI’s mistakes and shoddy output.
Agentic overload is another problem. When AI tools can generate vast amounts of output in minutes, the human reviewer becomes the bottleneck. The problem shifts from producing work to supervising it, and the load lands on the people least equipped to absorb it: senior partners who are already billing at capacity. There is emerging research suggesting that cognitive overload in law firms comes less from using AI than from supervising it, and we are only beginning to understand it.
None of these new problems replace the old ones. Clients still want better, faster, cheaper delivery of the same outcomes they have always wanted, and lawyers still need to draft, negotiate, and litigate. The new problems are real and deserve attention, but they sit alongside the old problems, not on top of them.
AI did not create the need for institutional intelligence. Deploying it exposed how much knowledge is still fragmented, tacit, or locked inside individual lawyers and disconnected systems. The cost of weak KM used to be deferrable more or less forever. A firm could run on reputation, heroics, and private form files and never be confronted about it. Buy a tool that is only as good as what you feed it, and the gap stops being theoretical.
If your KM program stalled in 2015, I would figure out why before buying anything else. Sponsorship, incentives, and getting someone to articulate and record why a document was deemed final are not technology problems. The new tools do not fix any of them, but they do make it much harder to pretend otherwise. That might be the most useful thing to happen to this field in a long time.
The tools are better than they have ever been, and the problems are the same as they ever were. The firms that name that plainly, instead of reaching for another rebrand, are the ones that will actually move forward. Maybe you don’t need an AI strategy. Maybe you need a knowledge strategy.
If these themes resonate, they are the same ones that come up repeatedly at the Knowledge Management and Innovation for Legal Conference (October 21–22, NYC). We built the agenda around them because they are the themes and issues with which the KM and innovation community is actually grappling.
–
About the Author: Patrick DiDomenico is Founder and CEO of LegalTech Connect and the author of Knowledge Management for Lawyers, published by the American Bar Association. He previously served as chief knowledge officer and chief innovation officer at major US law firms. He founded the KM&I for Legal Conference.
—
[ This is a sponsored thought leadership article by LegalTech Connect for Artificial Lawyer. ]
Discover more from Artificial Lawyer
Subscribe to get the latest posts sent to your email.