Updated: Aug 28, 2021
Future-of-law commentator Richard Susskind has written with Neville Eisenberg, a partner in Bryan Cave Leighton Paisner, this article in the current issue of The Practice. It’s entitled Vertically Integrated Legal Service, and it reflects a vision of the future of legal services that, when it comes to handling contracts, is at odds with my experience and what LegalSifter is doing. (I’m a lawyer who specializes in contact language; I’m author of A Manual of Style for Contract Drafting. I’m also chief content officer of LegalSifter.)
Here’s my summary of the article (using my words):
Little has changed in the legal market. Law firms continue to operate mostly as they always have, and alternative providers of legal services have yet to achieve massive growth.
The vision has been that routine aspects of complex transactions would be disaggregated—handled by alternative providers or new units in law firms and carried out at a much lower cost. That hasn’t happened. One reason is that disaggregation is challenging; another is that clients prefer keeping all their work under one roof.
This preference for single-sourcing presents a threat to alternative providers, as it has been assumed that the market would be segmented horizontally. It has been depicted as a pyramid, with alternative providers at the base, doing process-based work, and lawyers above.
But legal work cannot be treated purely as process and relegated to the lower reaches of the pyramid. You need a human lawyer in the loop—to check for anomalies, to control legal quality, and to resolve questions not in the playbook. That’s at the heart of reliable legal service.
The alternative is to organize legal work vertically by acknowledging that all legal work requires a mix of legal expertise, legal experience, and process. The law and lawyers should lie at the heart of legal service.
So one reason alternative providers haven’t scaled is that it’s impossible in practice and not desirable for most clients today for much work to be done without practicing lawyers being involved.
The obvious option here is that law firms themselves diversify and provide traditional service combined with lower-cost labor and technology, specifically artificial intelligence. AI would be used for, among other things, generating polished documents. These systems would be used by law firms and made available clients in the form of self-help systems.
All this suggests that for alternative providers to scale rapidly, they must recruit experienced lawyers. This presents an opportunity for law firms—if they can build vertically, supplementing their mainstream lawyers, they can compete with full-service alternative providers (where they are permitted) and dominate where alternative providers are prevented from offering legal services.
When You Forecast Change and Change Doesn’t Come
I haven’t read any of Susskind’s works. Or rather, I’ve read only just enough to have co-opted his handy coinage “legal knowledge engineer” (see for example this 2017 blog post). That’s because with contracts, the nature of the problem and what needs doing has always been sufficiently obvious that I’ve felt little need to look to others for inspiration. But I’m aware that since the 1990s, Susskind has been perhaps the name to drop for anyone involved in conversations about whether the business of law could be made more efficient.
For many of us, it has been frustrating that change has been painfully slow or seemingly absent. Susskind might have found it particularly galling; that might explain this new article, which sees change as not destabilizing the existing order but instead reinforcing it. Given Susskind’s reputation as an evangelist of change, to some of his readers that position might seem akin to apostasy.
Others will have their say on this article. I noticed this post on LinkedIn by Alex Hamilton, CEO of Radiant Law. In it, Alex takes issue with Susskind and Eisenberg by saying that putting law firms in charge of change won’t work, that it requires a very different kind of organization. He goes on to describe key attributes of Radiant Law.
Alex is the perfect person to address process. Me, I’m interested in who is best placed to do the required work, and the article raises two issues related to that.
Nonlawyers Can Be as Competent as Lawyers in Handling Contracts Work
First, Susskind and Eisenberg are lax in their analysis of the role lawyers play.
Let’s start by considering how they describe the work lawyers are needed for:
In daily practice, it is not possible simply to allocate everyday legal work to the lower reaches of the pyramid. A human lawyer is invariably needed in the loop—to check for anomalies, to control legal quality, and to resolve questions that are not in the playbook. This involvement of lawyers is not an occasional need. It is at the heart of reliable legal service.
Likewise, when large-scale deals are being conducted, the routine elements, such as document analysis and due diligence, cannot be handed over to administrative staff without supervision by human lawyers. It may be that this supervision consumes a small fraction of the overall time spent on document processing, but it is indispensable, high-value, and (for now at least), an irreducibly legal job.
In both instances, whether the work is everyday or bet-the-ranch, a simple horizontal disaggregation does not work. It is a pleasing theoretical model but, if put into practice, it neglects crucial ingredients—legal expertise and legal experience. It should come as no surprise that very few legal tasks can, in their entirety, be reduced to administration and parceled out to pure process providers.
That leads them to say this:
If it is accepted that legal expertise and, therefore, lawyers are needed for all legal jobs, the challenge for alternative legal providers becomes clear. Without lawyers or access to lawyers with appropriate knowledge and experience, their offering becomes of limited use—efficient process without sufficient legal assurance.
But Susskind and Eisenberg’s encapsulation of their argument, “legal expertise and, therefore, lawyers are needed for all legal jobs,” is an example of the logical fallacy “begging the question.” That’s because they base their conclusion on an assumption that is as much in need of proof as the conclusion itself. In effect, Susskind and Eisenberg say that lawyers are required to do work that requires lawyers. (Wikipedia offers as an example of begging the question Opium induces sleep because it has a soporific quality; for more on begging the question, go here and here.)
This sort of misfire spares you from looking at the nature of the work involved and whether it in fact requires lawyers. But Susskind and Eisenberg’s approach is unsurprising, because the same logical fallacy is built into the regulatory framework of the legal profession in the United States. Given the lack of a coherent definition of unauthorized practice of law (see for example page 2588 of this 2014 article by Deborah L. Rhode and Lucy Buford Ricca), the practice of law can end up meaning whatever lawyers happen to do.
We should be able to do better than that. My field is contracts, so let’s consider what makes sense in apportioning contracts work among lawyers and nonlawyers.
Quantum physics is a technical and specialized field—to do quantum physics, you have to be a quantum physicist. Working with contracts isn’t like that. Business contracts are business documents. There’s no basis for saying that only lawyers are equipped to handle the substance of business contracts—what you say in a contract. Businesspeople run deals all the time. That which is arguably the most “legal” part of business contracts—dispute-resolution provisions—is generally the part those involved in transactions are least interested in, with even transactional lawyers being inclined to wing it. (See this 2019 blog post for an example of that.)
Lawyers also shouldn’t have a lock on how you say in a contract whatever you want to say. Because we now have a comprehensive set of guidelines for clear and concise contract language (namely A Manual of Style for Contract Drafting), everyone who has some semantic acuity and is willing to apply themselves can be an informed consumer of contract language. Given the systemic dysfunction of traditional contract language, lawyers are in no position to argue that only they are qualified to be custodians of contract language.
Furthermore, with lawyers you’re paying for a credential that’s largely irrelevant to work with contracts. After all, in working with contracts I’ve applied hardly anything I learned in law school. I suspect that’s the case for most of us lawyers.
So it doesn’t matter what hat you wear; instead, what matters is whether, for purposes of whatever part of the contracts process you’re expected to handle, you’re familiar with what should be said in a contract and how it should be said. If nonlawyers have ground to make up, that can be accomplished with training and automation. I discuss that in this article from March 2021.
For me, this isn’t theoretical. This year I hired the first person to help me, in a supporting role, develop LegalSifter’s AI-plus-expertise technology. They aren’t a lawyer. I hired them because I’ve known them a while (we met through social media), they participated in my online training course, they’re possessed of semantic acuity, and they have considerable experience with contracts. Why would I reject them just because they’re not also a lawyer?
Frontloading Expertise Is More Effective Than Bringing in Lawyers for Remedial Work
My second issue relates to Susskind and Eisenberg’s assertion that “For alternative providers it is clear that to scale rapidly, to win major engagements, and to meet clients’ needs, it will be necessary to recruit a credible number of experienced lawyers.”
I can think of one alternative provider that appears to have the same view. This recent item in Artificial Lawyer, entitled LawGeex and the AI Gap, describes how LawGeex—like LegalSifter, it offers an AI-and-contract-review service—decided to supplement its AI offering with a human component. The article quotes a LawGeex person summarizing the situation:
“We started out as AI only, but saw that the tech was not mature enough. We took it upon ourselves to provide an end to end solution. And so we have lawyers [on staff].”
Furthermore, LawGeex has been authorized to operate in Utah as an alternative legal services provider. So they appear to be positioning themselves to compete with law firms in the manner recommended by Susskind and Eisenberg.
But what LawGeex has experienced isn’t so much an AI gap as an expertise gap. As I first observed in this 2018 post about a LawGeex marketing study, it appears that LawGeex has chosen not to invest in experts. Instead, they’re relying on nonexpert lawyers on staff.
That comes at a cost. Contracts are complicated, so to review draft contracts effectively, you have to wrestle with that complexity. If you’re applying AI to the contracts process, it makes sense to build expertise into the AI: you leverage the expertise, because all decisions the AI makes are better informed.
What is expertise? It’s being able to access all relevant information, filter out the noise, create order out of what remains, and make it accessible to the world. It’s different from, and rarer than, competence, which comes from applying to circumstances the guidance offered by expertise. Given the systemic dysfunction, serious expertise is required to make sense of business contracts.
If you don’t build expertise into AI, your only alternative is to add personnel at the back end, to make up for the limitations in your AI. And applying expertise at that stage is harder, and perhaps unrealistic. So for purposes of reviewing contracts, what Susskind and Eisenberg see as essential is in fact a remedial measure that is less efficient and might face challenges.
LegalSifter has made a different choice. We’re building expertise into our AI. For a glimpse of what that expertise looks like, see for example this video about notices provisions and this blog post about one of our jurisdiction “Sifters”—that’s what we call our algorithms. Currently the expertise is mostly mine, but we’re increasingly consulting other subject-matter experts. What we’re not doing is hiring squads of lawyers.
Building expertise into our AI is challenging, but the payoff is that it’s realistic to expect that with our AI looking over their shoulder, our users will be in a position make their own decisions regarding what’s in a draft they’re reviewing—they can take contract review across the finish line themselves. That approach offers our users control, quality, speed, and cost-effectiveness.
I imagine using the same approach to build my holy grail of contract-process automation, a comprehensive library of customizable automated contract templates. Technology would make accessible to many the work of a modest team of specialists tapping into subject-matter experts as necessary.
If an alternative provider wishes to compete directly with law firms by handling all aspects of contracts work, it would presumably need personnel to work with clients. If only to avoid unauthorized-practice-of-law problems, those individuals would likely be lawyers. But that goes beyond the disaggregation scenarios that Susskind and Eisenberg address in their article.
No Crisis of Faith
Because Susskind has for many years predicted a new order in legal services, it’s poignant to see him come round to the notion that law firms or squads of lawyers should be at the center of things. It suggests the quip that a conservative is a liberal who has been mugged by reality.