Metamorphosis: AI and the law firm
Integrating Generative AI into law firms will make the firms more efficient and their lawyers more creative. It will also completely change those firms' productivity, profitability, and purpose.
Earlier this week, I delivered the keynote address to LawDroid’s AI Conference, a day-long online learning event with an incredible lineup of speakers. After giving my presentation and answering some questions, I stuck around for as many of the other sessions as I could. I came away thinking that I’ve probably been too conservative in my projections for AI’s impact on the legal sector.
The first-order change, which we can all see from here, is that Gen AI will enable individual lawyers to carry out tasks much faster (and, if this research is any indication, to a higher degree of quality). The “individual” part is important. Generative AI is “an amazing personal productivity tool,” says one BigLaw CIO, and he’s right — but he adds: “We want to leverage that in overall operations of the firm and overall delivery of legal services from the firm.” I’m less certain that that’s where this particular road is taking us. I think “leverage” is not a word we’ll be using much anymore.
Gen AI seems best understood and deployed as a personal assistant that supercharges a worker’s productivity. I don’t mean “productivity” in the twisted, backbreaking way that law firms use it (“Our most productive lawyer billed 3,000 hours last year, and is now in intensive care”), but in traditional economic terms, as the ability to generate more output with less input. A machine that helps a lawyer complete an hour-long task in one minute obviously makes the lawyer more productive — that’s the “less input” side.
But Gen AI can also increase a lawyer’s productivity by enhancing the lawyer’s creativity in order to better serve clients — that’s the “more output” side. This creativity boost can come about both directly and indirectly:
Directly, Gen AI can produce new ideas and perspectives when employed as a suggestion generator, sounding board, and role-playing sidekick — like an incredibly inventive legal intern with boundless information and energy.
Indirectly, Gen AI can enhance the lawyer’s own creativity by giving it room to breathe — by unburdening the lawyer of grinding, repetitive, low-value tasks and freeing the lawyer to sit up straight, take a deep breath, look around, and think.
Now, efficiency scales, but creativity doesn’t. A law firm could implement a powerful Generative AI system throughout the enterprise (many are in the process of doing just that), potentially increasing efficiency tenfold or much more. But while it will become more efficient, the firm itself won’t become more creative — only its people will. Creativity is human, not organizational. Creativity can’t be leveraged. It resists centralized management.
But there’s a bigger problem here for law firms. By integrating into their business a new technology that improves both efficiency and creativity, they are playing against their own strengths.
Law firms don’t make money by being efficient — no business that sells its work and compensates its workers on the basis of time and effort has any interest in efficiency. Law firms make money by having lawyers perform legal tasks thoroughly and painstakingly, rewarding those lawyers for volume of work completed rather than the quality or impact of results. And in particular, firms profit from junior lawyers repeatedly carrying out low-to-mid-level tasks that half of them could probably do in their sleep at this point.
Law firms also don’t make money by being creative. Their lawyers often solve problems for clients through creativity, or create client opportunities, or add client value. But the application of creativity to clients’ issues happens relatively rarely — much legal work is rote (especially the leveraged stuff) and doesn’t call for ingenious imagination or fresh perspective. Senior lawyers and experienced advisors use their creativity more often, but their minute of creative insight is worth 1/60th of their hour of travelling to a deposition.
Law firms sell human effort; Generative AI vastly reduces it. Law firms neither reward nor make much use of human creativity; Generative AI kicks it into a higher gear. The more use law firms make of Generative AI, the less like law firms they’ll become.
Personally, I think that would be a fantastic outcome. Given the choice between a workplace that values the sweat of my brow and one that values my imagination and inventiveness, obviously I’m taking the second one, and you would too. Given the choice between a legal supplier that charges me for every hour worked by every lawyer assigned to my case, and one that charges me for solving my problems and facilitating opportunities and making my business or my life better, obviously I’m taking the second one, and you would too.
But I’m not sure those law firms that (rightly) are looking to make widespread use of Generative AI see the implications. By massively increasing lawyers’ productivity — by enhancing their efficiency and amplifying their creativity — they are forcing themselves to become productivity-based businesses. If they integrate AI into their operations, then they are drinking a potion that will turn them into something different — into highly productive enterprises that deliver value through lawyer creativity and that profit from systemic efficiency.
This is the second-order change that Generative AI, if it continues to evolve in its current direction, will bring to the legal services sector. It will render obsolete the organizing principles and business rationales of traditional law firms, and oblige lawyers to come up with new business entities — new value propositions to the market and to the workforce — that are better suited to post-AI realities. It will launch a sector-wide organizational metamorphosis.
This transition will not be painless. When presented with the prospect of a legal business that can’t sell hours, a remarkable number of lawyers I speak with run into a mental block. They honestly can’t see how they can make a living any other way. More than a decade ago, I wrote a quirky little sci-fi piece about a world in which time could no longer be measured, and how that would force law firms to change not just how they charge, but what they do. Generative AI moves us surprisingly close to that scenario.
I have a feeling I’ll be devoting many future issues of this newsletter to the increasingly pressing question: “How can we develop an AI-powered law firm that generates value and profits from the creative, efficient productivity of its highly skilled workers?” I can’t answer that question today — I simply don’t know — but I can give you an outline of what I think will be that law firm’s salient features:
Relationships with clients will be long-term and continuous, not episodic and irregular.
Relationships with workers (not exclusively lawyers) will be empowering and cultivating, not extractive and exploitative.
Internal collaboration will be essential to client-focused productivity, not a noisy distraction from hours-focused “productivity.”
Fees will be based on client relationships and strongly influenced by shared risks and jointly sought-after outcomes.
Customized training and professional development of workers will start on Day One and continue until the worker leaves the firm.
Most of all, a law firm that allows itself to be transformed by Generative AI (and the other macro-level forces at work in our world) will become more deeply attuned to the professional and ethical fulfillment of its clients’ goals. When you can no longer sell the time it takes to achieve a client’s outcome, then you must sell the outcome itself and the client’s experience of getting there. That completely changes the dynamics of what law firms are all about.
Back in 2017, I wrote: “A traditional law firm exists to provide buyers with access to solutions for their law-related challenges through the application of a lawyer’s time and effort. The future law firm will answer to the same description, minus the last nine words.” I was early, as I frequently am, but I don’t think I was wrong.