The unbundling of lawyer institutions
AI will strip law firms and law schools of their commodity features. Their future depends on whether they can rebuild around their highest-value functions and their trust-bearing core.
Two very different articles — one from a law professor, one from a legal technology analyst — crossed my desk last month. They each say something really important about law schools and law firms, respectively. But taken together, they point us towards what I think is an even more profound reality about lawyer institutions in the post-AI world.
At his eponymous Substack, Professor Michael Plaxton’s “To Our Next Law Dean” is really addressed to every dean of every law school, asking: After AI, how will you justify our existence? His concern is that AI is rapidly learning to perform many of the tasks law schools train students to do, and to deliver much of the general legal knowledge law schools provide at scale, including research, writing, analysis, and explanation.
But the high-impact services people really need from human lawyers — judgment, relationships, accountability, local knowledge, and contextual translation — are much harder to teach in a mass-learning context. So if machines can do much of what law schools teach, and law schools can’t easily teach much of what machines can’t do, then what are law schools for?
At Legal Technology Hub, Nikki Shaver’s “Law Firms Want to Change; They Just Can’t” asks whether law firms are capable of managing the transition to a post-AI legal market. It’s great that so many law firms are investing in AI; but the problem is that law firms are constructed — economically, operationally, and institutionally — around lawyer time and effort and matter-based delivery. Partner profit, associate advancement, and hourly billing all assume that legal value is created through lawyer labour applied to individual matters.
But AI changes the shape of legal work: It converts one-off tasks that once required hours of human effort into repeatable workflows that operate quickly, continuously, and at very low marginal cost. These new forms of value delivery just don’t fit with the structure and culture of the traditional firm. So even though law firms understand perfectly why they need to change in response to AI, they might not be able to.
Both essays are terrific, and I agree with their premises and conclusions. But I want to highlight an under-appreciated impact of AI on the legal sector that reading these articles together made me realize. Law schools and law firms are the legal profession’s most important institutions. But they were built for a world in which legal intelligence was scarce, and that world is rapidly passing away.
Legal intelligence — the capacity to apply legal knowledge to analyze legal issues, create legal instruments, and solve legal problems — is valuable not only because of its utility, but also because of its scarcity. Back in first-year law school, when our professors promised they’d teach us to “think like a lawyer,” they were also saying we’d become members of the exclusive class of people equipped and authorized to exercise that capacity — and therefore, one of the few reliable conduits through which legal assistance could flow. That rare ability would command a premium price.
Equally valuable, therefore, was any institution that could organize scarce legal knowledge and transmit scarce legal intelligence to novice learners at scale — which is where law schools have traditionally derived much of their economic value and their ability to charge what they do. And so was any institution that could organize the work of credentialed lawyers to solve legal problems and deliver legal outcomes to clients at scale, which is where law firms derive much of their economic value and their ability to charge eye-watering fees.
But scarcity was the key. In current parlance, the average person’s inability to organize and command legal intelligence to solve their problems enabled the “moat” that protected law schools and law firms. Accordingly, if legal intelligence were to become widely accessible — if, say, a machine could help learners acquire legal intelligence, or could mimic it to provide legal information, documentation, and direction to legal solution seekers — then both law firms and law schools would have a problem. And here we are.
AI can’t do everything law schools and law firms can. But it does change the status of generalized legal intelligence — the capacity for legal knowledge, research, issue-spotting, analysis, summarizing, and document drafting — from narrowly available to widely accessible. In the process, it strips value from every aspect of law schools and law firms that relies on the scarcity of legal intelligence. As it turns out, that’s quite a lot.
For law firms, of course, this means billable labour — the hours of lawyer effort consumed by those legal intelligence tasks that are done pretty much the same way across every law firm, and that occupy most lawyers’ time and generate most law firms’ revenue. (And it especially means the leveraged labour of non-equity lawyers, whose time is billed to clients with an additional profitable premium above their costs.)
And for law schools, it means the scalable delivery of general legal knowledge and instruction — the large-classroom transmission of essential legal concepts, principles, doctrines, and analytical pathways, all of which are transmitted pretty much the same way across every law school.
It’s important to understand: These aspects of law schools and law firms have always hewed uncomfortably close to commodity status. Their value didn’t reside in the unique nature of the curricular content or the legal deliverable — it’s not like the typical first-year Contracts course or the standard commercial lease was a bespoke work of legal art. Their value resided in the institution’s restricted or exclusive capacity to generate that content or deliverable, which was derived from the scarcity of the legal intelligence that produced it.
Now AI is turning the scarcity of legal intelligence into abundance. In the process, it’s exposing the commodity features of law schools and law firms. Those elements are now at serious risk — and as a result, so is the structural integrity of these institutions.
To be clear, law schools and law firms both perform a very wide range of roles and purposes that go beyond the narrow output of their legal intelligence work. Law schools provide students with not only intellectual development, but also identity formation and professional socialization. They also give the legal sector academic scholarship, original research, and sources of independent legal authority.
Similarly, law firms supply clients with legal products and services, but also with reputational assurance and trusted judgment in high-stakes matters. They also provide lawyers with gainful employment, novices with professional formation, and the community with centralized access to (admittedly expensive) legal assistance.
In other words, law schools and law firms are bundling mechanisms — they gather together myriad important functions within a single institution, some differentiated and some not. These institutions are large, and they contain multitudes. But AI is now leading a wave of commoditization that will wash over everything legal-intelligence-related in both law firms and law schools.
In the result, these institutions are going to experience what rocket scientists might call “rapid unscheduled disassembly,” but what we would call unbundling.
Unbundling, for our purposes, is the process by which all the disparate elements of a complex process or institution are isolated and examined to see which could be more usefully performed or acquired elsewhere. So the question becomes, “What can I get from an AI, adequately and affordably, that once I could get only from a law school / law firm?” The answers will include more than schools or firms might like to admit.
Those answers would certainly include expert instruction in and effective assessment of legal knowledge acquisition. As Michael Plaxton says, “The required courses in law school have a shared core content … that renders them more or less interchangeable across schools. … This interchangeability and generality of much of what we teach suggests that AI can do it at least as well — and faster and cheaper.” I’m confident that a frontier AI system could teach me American antitrust law as effectively as an American law professor could — more effectively, in fact, since I would be the AI’s only student and I would have 100% of its time and attention to teach, test, evaluate, retest, and so on.
And the answers would certainly include basic and intermediate legal information, documentation, and direction — assets that law firms sell by the lawyer-performed hour, at rates which still presuppose an exclusivity that increasingly will no longer apply. And as Nikki Shaver points out, the “lawyer task performance hour” is more than a pricing mechanism: It’s the atomic unit of law firm structure, finance, and culture. Removing the lawyer-performed billable task from the traditional law firm is like removing the most important strands of a person’s DNA.
What if we turn this question around and ask: What is not commoditizable about a law school or law firm? Again, the answers are narrower than we might think.
A law firm’s truly unique or differentiated assets would include its very best lawyers, those with exceptional judgment, credibility, and relationships — assets that few other lawyers possess and no system or machine could replicate. They would also include a powerful reputation or brand that tells not just clients, but also shareholders, regulators and others, that a decision to rely on this firm’s advice and direction is safe and justifiable. A truly differentiated law firm says: “You can rely on our judgment, and even if that judgment fails, your decision to rely on us will be defensible.”
Interestingly, a similar dynamic operates for law schools. A school’s most valuable asset is also its reputation, which communicates not only the quality of the instruction its graduates received, but also the legitimacy of the certification process through which they passed. Sure, some of that reputation can be attributed to the host university; but at least as much can be attributed to current or former faculty members recognized as occupying the peak of their academic disciplines. A truly differentiated law school, therefore, says: “You can rely on our graduates, and our rigour and reputation will make your hiring decision sensible and secure.”
In both law firms and law schools, therefore, the scarcest and most valuable assets are great people and great reputations. In a real sense, each of these assets enables and reinforces the other. And what both these assets share in common is one thing: They engender trust. And now we come to the heart of the matter.
A functioning society needs reliable repositories of trust. We don’t all know each other and we don’t necessarily all like each other, so we need recognized institutions in which we can collectively deposit enough trust to enable our daily transactions. We need repositories whose foundations and standards and safeguards are so strong that they justify our confidence; we can count on them and on what they provide. Houses need load-bearing walls; societies need trust-bearing institutions.
Law schools and law firms have always been the legal profession’s trust-bearing institutions. AI is going to unbundle each of their individual functions and deliverables, stripping the value from any whose commodity status AI has revealed or enabled. That might sound like a pretty dire forecast.
There’s more at work here than the economics of efficiency, however. The human need for trusted legal institutions is real, and it will survive the impact of AI. But these institutions have to recognize and accept the surgery they’re about to undergo — the removal of their commodity elements — and make plans for how they will recover, adapt, and flourish afterwards.
Law firms bundled elite judgment with leveraged labour and institutional accountability. AI will remove much of the labour, but the need for judgment and assurance will remain. The law firms that survive this process will be those that retain or develop elite judgment and accountability and reconfigure their financial and cultural infrastructure in order to profitably sell the judgment and accountability. As Nikki Shaver says, that is an extremely tall order. AI-native law firms already have a head start in this race. But the race has only begun.
Law schools, similarly, bundled academic rigour and integrity with scaled legal instruction and institutional gravitas. AI will remove much of the scaled instruction, but the need for rigour, integrity, and gravitas will remain. The law schools that survive this process will take to heart Michael Plaxton’s observation that in the aftermath of mass instruction, local and customizable professional formation becomes invaluable. Nobody knows where the next generation of experienced lawyers will come from — but nothing says they can’t come from law schools. Nobody is preventing law schools from becoming customized lawyer formation facilitators. That would be an even taller order than what law firms are facing. But it is possible.
The bottom line is that society needs lawyers it can trust, and I choose to believe that it always will. We will need lawyers who can advise, advocate for, and accompany clients on their most important legal journeys; who can stand up for the rule of law, hold power to account, and expand the reach of civil justice; who can envision and launch Legal Moonshots that use the law to change the world for the better.
For everything that AI can’t do — and for everything that AI will do that we’re really not going to like — we will need lawyers who have been formed with rigour and integrity, and who can serve clients with judgment and accountability.
Many law schools and law firms filled that role in the past. Some can fill it again in the future. But only if they understand what AI is about to do to them, if they surrender the commodity functions whose scarcity value is disappearing, and if they rebuild themselves around their highest remaining purpose: becoming the trust-bearing lawyer institutions the post-AI world will need.
Jordan, I think you're onto something important here, especially in your emphasis on trust.
Reading this, I found myself returning to something that has been reverberating in my mind a lot lately. I tend to think less about the institutions and more about the people inside them. If legal intelligence becomes increasingly abundant, then many of the things that brought lawyers status and value for decades are going to be viewed differently. Not disappear, but occupy a different place than they once did.
In particular, my perception is that most of the discussion around AI in law seems to focus on technology and business models. Those are obviously important. But I wonder whether legal leadership is underestimating the personal side of the transition.
Many senior lawyers built careers around intelligence, expertise, analytical rigor, and mastery of information. Those qualities deserve respect. They also become part of identity. A lawyer can fully understand the need for change and still struggle to adapt to it. In my experience, the obstacle is often not disagreement with the future. It is the challenge of letting go of a version of oneself that has worked very well for a very long time.
Lawyers are particularly interesting in this regard because the profession trains people to analyze, explain, and argue. Those skills are enormously valuable. They can also make it easier to intellectualize a transition than to live through it.
Your article focuses on trust. I agree that trust becomes more important, not less. But trust ultimately gets carried by individual lawyers. The profession already struggles with anxiety, burnout, and distress. It seems likely that a transition of this magnitude will create additional pressure, particularly among the people who have invested the most in succeeding under the existing model. When the ground beneath you does not seem stable, you trust yourself less, and it becomes increasingly difficult to exude trust.
That strikes me less as a wellness issue than a leadership issue.
I am working on a manifesto for AI in our human-centred law firm and your excellent piece resonated. Only last week we gathered our partners to look at this and with care to ensure our culture and what makes us who we are is supported not supplanted by AI. A reliance on empathy, judgment and trust is what will always endure in my opinion. And…it won’t be easy. Your piece speaks to the fundamental redesign for many law firms. High staff costs and fixed overheads will feel even more cumbersome as some supertankers seek to turn, to downsize and set new courses.
Great read. Thanks Jordan.