About a month ago, Steven Croley, Chief Policy Officer and GC at Ford Motor–who, full disclosure, I do not know–published a column on Bloomberg Law opening with the commonplace observation that AI is disrupting the legal services market for large public companies such as Ford. But then he immediately posed the question which has been front of mind for us here at Adam Smith, Esq. for some time: Will outside law firms or in-house counsel harness AI “most effectively,” meaning not as a super-sophisticated tool–think spell-check powered by a rack of servers–but “as a surrogate legal service provider.”
I hope we can stipulate at this juncture that the legal services marketplace is past the brink of rapid change and in the thick of it. That has specific and pointed implications for law firms, namely that ongoing in-house innovation driven by AI will raise expectations for comparable or superior AI-powered innovation from outside counsel.
The two sectors are, to be sure, responding to different dynamics: In-house teams are closer to their clients than outside firms ever could be, but the flip side is that firms are exposed to a far greater array of clients and a broader span of the legal landscape, “degrees of difficulty” in issues posed to them, and wider array of expertise across their partners and associates than a focused single-client driven inhouse department.
So to pose the question pregnant in these observations a bit more directly: Think of AI as constituting a brand new third legal service provider for clients, in addition to the corporate in-house legal department and outside counsel. Few concepts in economics are more fundamental than market structure. The market structure of legal service providers has changed. This is not a cycle; it’s never going back. What economic and performance advantages does AI bring and what disadvantages? And since technological progress is relentless and unidirectional (Moore’s Law), how will AI’s span of superior expertise for the money evolve?
Already today, LLMs are masters of complex research–their bots have scraped the entire internet and forgotten absolutely nothing, and their built-in unceasing and recursive pattern of behavior means they’re almost perfectly up to date. More importantly for the legal domain, they can produce highly sophisticated and essentially immediate summaries of their research into any plausible question posed to them, in grammatically correct and fluently composed paragraphs suitable for copying and pasting as is into client-ready material.
(If you haven’t experimented with this yourself, I can hardly recommend it strongly enough; it brings new meaning to “uncanny.”)
The skeptics in the audience are scoffing at the practical and professional risks (malpractice? yes, we’re seeing it already, with consequences all the way up to threatened disbarment) of unvetted AI slop passed off as work product. That is emphatically not what I’m envisioning.
Rather, where GenAI fits into today’s legal workflow is in the position of a talented senior associate; it gives the partner high-caliber raw material to work with, and to massage with their established internal grey-matter library of legal experience and EQ skills at client relations and client management–the famous, if tenuously grounded, 10,000 hour rule.
So far we’ve been dwelling in essentially the tech world perspective on AI: We’ve been asking what can it realistically do?
Now, to the economics of the topic. I hope you’ve anticipated what I’m about to say because I think it’s lay-down obvious: In this world, the billable hour is fading fast–it’s less and less “fit for purpose”–and so is its revenue-enhancing rocket fuel, the pyramidal human capital leverage model of staffing matters. Law firms will have to design, and put into practice, an entirely different revenue model. The only plausible candidate I can think of is to charge by the project (or the project phase for large engagements). This will require law firms to build an entirely new set of financial muscles, and there will be potholes and money-losing engagements along the way.
Even though this will be a novel experience for firms, there’s no reason–mismanagement rising to the level of malpractice aside–to think it’s a threat to law firms’ futures. That may sound like a once-in-a-career change, and it will be, but it’s by no means the least of it. Still, we’ll get through it. But that’s not the only change afoot.
Step back and look at the structure of legal services supply and demand using a clean sheet of paper, now including the services provided by agentic AI.
Previously we have had the familiar bilateral market, with clients providing the demand and law firms the supply. Two dimensional, if you will.
We’re leaping into a multi-dimensional market structure with the introduction of commercially viable and high performance agentic AI. Do not be tempted to fall into the intellectually lazy assumption that agentic AI is merely another arm’s-length provider of legal services to corporations, something Ford Motor and everyone else can purchase as an alternative to or arm’s length substitute for the AmLaw.
Law firms seem to have warmly embraced this view of things. Ford and Microsoft collaborated on a survey on this topic, with an anonymized (no attribution) section where law firms were asked to rate their own level of maturity on AI adoption as well as their competitors and their clients–and their clients’ expectations. Here’s where it gets interesting, or amusing. At a simple qualitative level, over two-thirds of firms confidently assure clients that they deploy GenAI at a level with real commercial impact. But this was mere assertion, with no case studies or examples. And indeed, reason for skepticism came from the answers from the one-third who provided real examples in their responses, where the successes were extremely narrow and thin.
There’s more.
Seventy-four percent of law firms rated themselves as further along than their clients (94% three years hence). Yet only 30% thought law firms in general are living up to what clients expect, today or three years out. Who are these lagging firms? Well, it’s not the survey respondents: 69% reported confidently that they exceed clients’ expectations today and they see that climbing to 90% within three years. Obviously, the math doesn’t work: 70% of the industry falls short of clients’ expectations–or else 90% exceed them.
But let’s come back to Ford. Law firm leaders take comfort in believing that Ford is an outlier in judging its internal legal capabilities superior to law firms–a sport, if you will, by no means representative of Fortune 500 clients. I’m in no position to judge whether those leaders’ comforting beliefs are accurate or not, but I know a lot about the evolution of markets over time, and when massively disruptive new technology is introduced, the complacent incumbents who rest on the sidelines will not be around to see how it plays out.
We’ll give the last word to Croley at Ford:
“We will continue to invest and expand our AI capabilities in-house, even as we solicit innovation from external legal service providers and are prepared to pay for demonstrable gains in productivity. We are likely to expand our relationships with external partners who innovate and deliver quickly and curtail our reliance on those who lag behind.”
Ford Assembly Line as envisioned by Gemini AI
