[author: Rick Jones]
I think about AI off and on in a desultory sort of way. I wonder about its threats and promises; a sci-fi dystopia of the Industrial Revolution 2.0. One thing I share with our most recent wannabe president, Ms. Harris, I, too, know that AI stands for Artificial Intelligence.
I also suspect like almost everyone, that pretty much sums up what I know about AI.
I’ve listened to many efforts to explain AI. I’ve gotten to the point where I can nod knowingly at a cocktail party about its claims, prospects, psychoses and threats to life on earth, but that’s intellectually pretty thin gruel. It’s clearly growing daily in scale, power and ubiquity (and no, I didn’t use AI to write those words) and I am convinced that its ramifications for the culture and for the business landscape, will be huge. (I really have no clue, but for purposes of this commentary, I’m going with huge.).
I’m also convinced that, like most profound changes, much of what will be wrought, will remain unanticipated and not understood until after it happens (but that’s not going to stop me from speculating).
I recently heard a very thoughtful and well-respected observer of all things geopolitical describe AI as the biggest and most consequential emerging change in (threat to?) the human condition that will have occurred in generations. That’s pretty astonishing considering the competitive set of things that might compete for the title of Most Consequential Event including war in Europe, war in Asia, war in the Middle East, pandemics, economic and political disarray in the United States, zombie apocalypse and the possibility that the Jets will win a Super Bowl. That brought me up short.
There’s a certain seductiveness to turning a blind eye to the likely impact of AI. It’s too big, too complex and full of technical innovation which might as well be magic to me. Nonetheless, doing my part as a board-certified talking head, I have thought a bit about how it might impact my old patch, Big Law. My guess is that it will do many, many things in and to Big Law but for the moment, I want to focus on two very foundational things: law firm training and law firm pricing. These things will be deeply intertwined in a dance macabre when AI remakes the business model.
Let’s start with training. What do we do with all those junior lawyers who, as all acknowledge, start off their career not really knowing where the restrooms are, yet are paid a great deal of money in that blissful innocence? The hoary old training model is to have young lawyers do relatively rote and ministerial work (note: it’s not rote and ministerial the first time several time one does it) and in the doing, and concomitant exposure to more senior supervisory lawyers, learn the craft and grow in their understanding of the law. Eventually they become (one hopes) real lawyers.
AI will require a rethink. If an AI program can generate first, second and perhaps even third drafts of key transactional documents, write early drafts of briefs and motions and provide detailed suggestions to the humanware about how the documents or arguments could be changed in response to new facts or in response to counterparty comments, one must begin to wonder what all those young lawyers will actually do and how they will learn to think like…well, lawyers. Now, not to overstate this, I don’t care how good the AI gets (I hope), there will be a role for lawyers to craft arguments, write complex language, intuit the real needs and incentives of counterparties, determine what’s important, work with the other side of a transaction or litigation to find common ground, etc., and some of that work would be appropriate for young lawyers, but will there be enough of it to sustain our current approach to the business of law?
Many worry, and I include myself in this “many,” that AI will change how we think…and not in a good way. In a recent excellent article in the Wall Street Journal Allysia Finley, writing about AI, observed that “the brain cannot draw connections between ideas that are not there.” Hmm. That’s a problem for anyone and surely for aspiring lawyers. If we increasingly rely on AI to build arguments, articulate conclusions, suggest connections between disparate facts (based upon the tools’ billions of data points and not our own actual experience), does our ability to think in the ways that lawyers need to think atrophy? I have always thought that the characteristics that make a lawyer particularly good was a driving curiosity, the ability to see through the superficial to the underlying superstructure, to intuit the underlying motives and incentives of all parties to a matter and the ability to connect disparate facts and see patterns where others do not. If we abandon that skill set to the machine and stop wondering, stop musing about what might be really going on in the matter, about the knowns and unknown unknowns, these capacities will likely degrade.
I’ve heard all the happy talk about how this will free up young lawyers to think, ponder and to engage in deep analysis with less time devoted to the rote. This seems like magical thinking. One needs stuff to inform analysis. Analysis doesn’t exist independent of the “facts” (educators, please take note). One cannot make cogent arguments about the democratic form of government without some historical knowledge of our Revolution, about Locke, Burke, Rousseau, etc. You know, the data…the dots. If you don’t own the dots, you won’t be able to connect the dots, you won’t see the relationships. If you rely on AI to tease out all these relationships, you won’t develop the skills to do so and will not indeed become a good lawyer.
We’re going to have to materially change the training model. We’ll need to recognize that very junior lawyers’ value proposition will be increasingly degraded and find a way to still produce great lawyers. Frankly, I suspect that we’ll discover that one to three year qualified lawyers will soon have very diminished economic value, but the only way you get senior lawyers and partners is to train up these youngsters.
Training must no longer be the haphazard drop-them-in-the-deep-end-of-the-pool approach endemic to Big Law learning while billing 2000+ hours. These youngsters simply won’t have marketable skill sets in the first few years after leaving law school. The clients won’t pay for it and as AI is more comprehensively deployed, there simply won’t be the hours. Perhaps we need to embrace the UK model where young lawyers, after law school, spend two years at a law firm as “trainees” learning their craft. During their two-year training contracts, the law firms utilize the talents of these young lawyers to a certain extent but also teach them the business with a formal educational and grading system imposed by the Law Society on the law firms themselves. (Note: It would be best if our law school friends could be part of the solution. I envision that, ultimately, the academic segment of the law school slog will be shortened and young lawyers will receive more of an apprentice-like experience in the first several years of practice.)
Maybe for our trainees we turn off or limit access to the AI Deus Ex Machina? We make our young lawyers reason based upon what they know and what they can learn through traditional research. We make them look for patterns and not ask a machine to provide the patterns. Force the habit of thinking before we unleash the technology?
Look, if AI does everything that it has currently promised or threatened, we simply will not be able to build young lawyers like we currently do. We need to fess up to that reality. We need to train them up for several years before they will have marketable skills and in the doing, they will have less billable time and when we bill them, we will almost certainly bill them out at lower billing rates (again, not inconsistent with the existing UK model). In consequence, we must pay these trainees less than the breathtaking compensation package they receive today if we want to preserve profitability.
Corporate America has been carping about the uselessness of junior associates and the relatively high billing rate for those associates for years, largely to little effect. The law firm pyramidal business model has remained a constant for decades and seems immutable (excepting, of course, for us obviously raising billing rates every year). Everyone complains, but nothing gets done; the financial equivalent of the weather. AI will change this.
An underlying reality of the big law business model is senior talent is priced too low and junior talent is priced too high. The clients agree with half of that statement. (I suspect you know which half.) In an alternate universe senior lawyers, who bring judgment, acumen, experience should probably be paid at a multiple of the current (yes, already high) billing rate for such lawyers and the billing rate for the junior lawyers would be a fraction of what’s currently charged.
However, it’s time to change the paradym entirely. The answer to the materially increased productivity of the practice of law and the dismantling of pay and pricing pyramid must be value billing. Value billing means what it says, the amount to be paid for legal service must be determined by the value the client received for that service and not based entirely on the cost of the lift.
There has always been talk in the law biz about a return (yes, I mean return, not innovation) to value based billing where the parties can agree on the value delivered by the legal lift and price that service accordingly. That was standard in the US legal market 100 years ago, and it’s still embraced by a handful of super-elite firms in at least the M&A patch, but the hour billing model has broadly prevailed. Oh, we still talk about value billing and value billing exists in some transactional practices, but it always comes back to the hour.
When in the past we have talked about alternate pricing methods, the law side has pointed out that there’s sometimes little connection between the amount of work required and the value of the underlying transaction (closing a billion dollar mortgage finance transaction and a $50 million mortgage often requires the same legal lift). (It doesn’t help that the biz side largely and consistently, wildly underestimates the amount of legal work to achieve their goal…but that’s another story.) Some on the biz side quite publicly but surficially embraced value-based billing, but let’s be clear that when client embraces value-based billing, what they really mean is value-based billing capped by an hour-based billing metric. No meeting of the minds here.
AI will cause all that to change.
Value-based billing must become the dominant pricing infrastructure of law firm economics. Here’s where AI (and I’m sort of talking through my hat here) can be helpful in assessing the cost of our performance (not the price!). Law firms have struggled with practice management schemes in the past trying to predict the legal lift associated with particular matters, but this effort has not been terribly successful. In the past, there’s simply been too much indeterminacy in a prospective matter, too many complexities to juggle, too little usable data to develop confidence. With the new technology in hand, perhaps we will know with more certainty the legal cost of a particular matter should or ought to be. Perhaps AI can fix that by crunching much more data, looking across many more matters, and juggling all the complex variables to come up with something that is reasonably reliable as to the cost of a matter. How many times have you sat around with your colleagues and said, “Gee, we did something like this before. Does anyone remember the name of that transaction? What did it cost? What were our write-offs and how were we staffing that transaction?” That conversation generally devolves into well, let’s go have a beer and hope the client will simply pay for what’s on the books. That’s going to have to change.
And let’s remember value billing cannot be a vogue-ish euphemism for charging the cost of performance. The nexus has to be broken. Like any manufacture of a product, it will always be critical for law firms to know what a requested performance will cost, and when talking costs here, costs much include IT, CapX IT staffing and the entire cost of the training module. Armed with good cost information, we can price our work. That will entail thoughtful conversations with the clients about how to measure the value of the legal performance. I don’t know how all this will work out, but I know it will. Will we develop some rules of thumb for pricing? I certainly hope so (I do remember long ago when people used to bid legal on mortgage finance transactions as a number of basis points against the amount of the loan). We’ll come up with something. If these conversations have to be de novo every time, we’ll all go crazy.
All this will change the law business entirely. Law firms are likely to shrink and get much more senior heavy as the traditional economic pyramid underlying the current law business is broken and AI is fully deployed to bring productivity and efficiency to a business which really has not been particularly interested in those things. The clients will still require consigliere-type lawyers. They will require brilliant transaction lawyers and brilliant litigators who can innovate and solve complicated problems. In military talk, the teeth will stay the same, the tail will be radically restructured.
Big Law is surely on the case. Most big law firms have been and continue to experiment with a wide range of AI tools.
What I’m less sure is happening is whether many of us have yet confronted the reality of the radical change embedded in AI. To change our training module and to change our pricing module is huge. Those are changes that will cascade through every bit and piece of our practices. Incremental is not going to get us there. We can’t tweak our way into this new world order. I know that as lawyers, we are largely suspicious of big ideas and big change and we are incrementalists at heart, but we’re at one of those inflection points where big thinking is required for big change.
When the combustion engine car showed up, you didn’t compete with a better bridle for your horse. Winners will be folks who figure out how to fully adjust to this change and fully embrace these new technologies.
Firms should, at a minimum, already have started to conceptualize what this new world might look like. Firms should model revised pricing and lawyer development approaches, indeed prepare the groundwork for radical change. This is change that might not need to be implemented this year or next year, but it will need to be fully implemented rather soon.
Transition will be painful. Until this new reality proliferates rather homogenously across our business, conversations with clients on the one hand and with talent on the other will be hard. I can barely imagine how hard it will be going to our favorite T14 law schools and suggesting to the young masters of the universe that we want to pay them a great deal less than recent law school graduates received whilst learning where the restrooms were and doing mountains of rote work. We’ll need to figure out how to incent the best and brightest to embrace this model (probably with a notion that there will be a substantial payoffs down the road). Many law firms will not adapt quickly (think Neanderthals) and so the old training, pricing, associate compensation and develop models will be running in parallel to the new. That will make implementing the new models harder but no less imperative. It’s going to be hard and no one will want to be the first mover. But if Big Law goes on hiring enormous cadres of first year lawyers and paying them enormous salaries in a world where the client base will demand to receive the benefits of AI productivity, profitability will suffer. We all know what happens when profitability begins to suffer. As Samuel Johnson famously said, “Depend upon it, Sir, when a man knows he is to be hanged in a fortnight, it concentrates the mind most wonderfully.” We’re there; time to concentrate.
And you thought all you had to worry about was war, disease, economic distress and the poor Jets.
