Loading...
The Legal Industry Has a Data Problem AI Can’t Fix
This episode segment explores how legal technology, data quality, and artificial intelligence are reshaping the legal profession. The speakers compare law firms’ relatively limited investment in data hygiene with the financial services industry’s regulatory-driven focus on data governance, anti-money laundering, and risk management. They discuss how legal teams can become more integrated into business operations by using AI to reduce administrative work and focus on outcomes. The segment also examines the risks of AI hallucinations, weak supervision, inadequate training, and the continuing accountability of lawyers.
Host: Hotpur, thanks for joining this afternoon.
Hotpur: The Owner’s Mind.
Host: And actually, I guess it’s evening your time, right?
Hotpur: Just hit 10 p.m.
Host: Oh my goodness. Well, I appreciate you carving out time, because that doesn’t sound like a lot of fun, to be sitting on the computer talking to me at 10 p.m. So thank you.
What do we start with? Just a quick intro—who you are, what you do, where you do it?
Hotpur: Sure. In brief, I’m a computer scientist at heart. I worked in the City at an investment bank, ended up as a lawyer, and then decided to figure out a way of combining the two together.
So technology has turned a lawyer, and I’m finally living the dream and putting them together.
Host: And I guess the environment is such that it’s a good time to be doing that.
Hotpur: It’s very true.
Host: You and I had a great conversation the last time. I spent years in financial services also, and it’s a very different world in terms of the focus on data hygiene and data quality.
People talk about legal being highly regulated. They have no idea about highly regulated until you’re in a bank. You’ve got the alphabet soup of regulators: FINRA, FNSA, and SEC, OCC, Federal Reserve. Those are just the big ones. Then you have local regulators.
It’s quite a—there’s a lot of scrutiny in financial services. I joined financial services right before 9/11, and shortly before Sarbanes-Oxley and the USA PATRIOT Act. That was a direct result of 9/11 and the amount of scrutiny around anti-money laundering.
I spent time in anti-money laundering. Actually, most of my time at Bank of America was in AML. The amount of focus on terrorist financing, which is a subset of AML—AML started really to bust up organized crime, and terrorist financing was a later development.
After 9/11, that became a top-three risk for financial institutions. We had to start paying really close attention to anything related to KYC, which is know your customer.
Quite honestly, the data wasn’t there to really be compliant with KYC. We had to do a lot of work. But we were still much further ahead than law firms are today.
Hotpur: You know what, I’m processing what you’re saying. Look, I was front-office IT, right? I can’t believe I was building pricing derivatives, quite frankly.
I felt we didn’t take data seriously either. It was actually the financial crisis that forced us to take data seriously.
Host: I completely agree in terms of the impact 9/11 had on AML and making sure that data hygiene was right. But the financial crisis brought it home to me.
It pushed these financial institutions, also through things such as BCBS 239, and forced data governance and data cleanliness where it really mattered and led to financial stability.
The other reason I was chuckling is that I teach at the University of Surrey. I teach financial technologies and financial regulation and fintech policy, which kind of go hand in hand with that.
My students are always amazed when I talk about some of the legislation that came in—the Terrorism Act, anti-terrorism, terrorist financing, et cetera.
I talk about Landsbanki, the Icelandic bank that collapsed, and when they couldn’t get deposits back to the savers in the UK or in Europe outside of Iceland, they pulled up terrorist legislation to say, “Give it back,” et cetera.
I’m just smiling because a lot of these things are now coming together, right? You talked about a highly regulated industry. It’s all starting to come together.
Hotpur: There are definitely those moments where you see it progressing.
I suppose you’re probably struggling to do one of these podcasts these days without talking about AI. Data is key as part of it, right? I think that’s going to be a big moment as well.
I just wonder whether, in podcasts in a decade’s time, people are going to be having that moment about law.
Host: Yeah. So I was there through 2008 as well and left shortly thereafter. The stress tests that the regulators were applying to financial institutions to ensure the OCC’s mandate—safety and soundness—were significant.
It’s interesting how Sarbanes-Oxley was a big wave of controls and data quality, and then there was another wave with the financial crisis.
We have not had that sort of event in legal yet.
I mean, we’ve had events, right? I don’t know if you remember the Panama Papers. I don’t know when that was. It seems like it was 10 years ago.
I think that was an awakening of risks—existential risks—to law firms. A breach could be a really big deal for a law firm.
But the blocking-and-tackling data that sits in the DMS, taking information security out of the system, just the data hygiene of it—there’s been relatively little consequence for law firms to not really invest there.
As a result, they largely haven’t. Now we’re in a place where there is a cost associated with the lack of data hygiene, and that’s AI, which is kind of what you’re talking about.
Hotpur: I think we need to look at it from the in-house angle, that has seen that, certainly within financial institutions.
I always thought, look, I started my journey into law in 2003. That’s when I had this wake-up moment. Reading some of the Scandinavian work and *The End of Lawyers*, et cetera, I felt that lawyers had to wake up and worry about the data.
But that was very much from the perspective of a bank needing to know what these things are—what these financial instruments are—and quite honestly, they’re just contractual obligations. So that brings legal into what running a bank needs as part of it.
In the financial crisis, what we saw is that all of a sudden, those terms that went in—and there was a particular term called rating downgrade clauses—actually mattered.
I remember being on the covenant at a bank, and I had the audacity, as far as the traders were concerned, to dig my heels in and stay in line with credit policy and say, “We can’t put this rating downgrade clause in.”
It basically said that if the bank gets downgraded so many notches, then the counterparty has the right to terminate, which sounds reasonable, right? It’s a material change in the creditworthiness of who you’re facing, and therefore the whole commercial deal is a different one.
I’ll never forget that the trader pulled me aside and said, “What the hell are you acting for? If this happens, I’m not there, you’re not there. None of us are there. Who cares?”
It’s really interesting to me that you fast-forward. The financial crisis occurs, regulation comes in—a lot of regulation comes in—and you’re now asked to stress-test.
You’re asked, “What if there were a three-notch downgrade under these clauses? So the bank’s creditworthiness is squeezed out. We want you to tell us how much money goes out of the door as a result, right? What is the financial impact? And we want you to take a percentage of that and set that aside as regulatory capital.”
All of a sudden, that clause that someone said, “Well, unless it happens, at which point the end of the world is here as far as we’re concerned, so we just don’t need to worry, because we’re not around at that point,” all of a sudden gets pushed onto the in-house legal team.
“We need to know about this because it’s got a consequence—not now, not when the downgrade or some nuclear event happens, but it’s got an impact now, because we’re forced to say, ‘What if the world’s conditions change?’”
We don’t want another financial crisis.
I saw in-house teams having to worry about that. Coming back to what you were saying, the problem is that I think law firms neglected that.
The lawyers were one step removed because they were just solving the legal problem of documenting these things and not feeling that if you could understand what was in these agreements and merge it with transaction and trade data, then it gave it meaning.
Hotpur: I remember being back at Allen & Overy, and we had finished a big securitization deal. I found myself being an expert on derivatives in the context of securitizations—a niche of a niche area in many ways.
There was this ridiculous—or, for what it was, ridiculous, but very important—form that the client’s in-house legal team needed to fill out. Frankly, it was a very thorough summary of the terms that I, as external counsel, was putting into the documents for them.
I remember talking to this in-house lawyer and just saying to him, “Well, why don’t I fill it out?”
He said, “What do you mean?”
I said, “Well, you’re being asked to put it into a system on your side. I’ve just drafted all of these documents. I know them in my sleep, and if I don’t know them in my sleep, then you’ve got to probably just send me the questions. Send me a screenshot, and as part of what we are providing, I’ll get it to you.”
It was amazing how quickly that in-house lawyer gave it to me, because they clearly hated to do it.
But it was all about data by that point, right? It meant so much. I think it’s actually quite sad for me that it’s taken so long for law firms to recognize that, and it’s taken technology and AI to mature to suddenly have its moment.
The whole world—my children know about AI. Find me a person that doesn’t know about AI at the moment.
Again, so many people that you speak to, lawyers ask, “When did AI happen?”
“2020, 2022,” et cetera.
Host: When ChatGPT became public.
Hotpur: No, it had always been there for a long time, right? And the need for data has been there for a very long time.
It’s just that law firms haven’t seen it. They’ve worried about solving legal problems.
I actually think that’s meant that they’ve not listened to their clients—the people they work for—that have been saying the digital agenda hasn’t just arrived.
Particularly in a bank, you don’t just walk up with wads of cash and pass it over. You don’t just do things manually. There have been systems and data for a very long time.
But as lawyers, we’ve chosen to say, “We’ll sit in our ivory towers and continue working through words and words only, and not data.”
Host: I had a conversation with Bhavna Tilman, who was on the pod. She’s actually on this week’s podcast. Our episode will air in a couple of weeks. We’re recording on August 13th.
That’s a great episode.
The way Bhavna described the legal department of the future was very different from what I experienced in my years in corporate America.
As part of my AML work, we had to engage with legal frequently, as you can imagine. There was an absolute firewall between us.
I was in risk management, which is not in the business. It’s a risk management function aligned to various areas of the business, and we had to go through a firewall portal. We didn’t have direct access to legal.
Legal even had a separate reporting structure. It maybe had a dotted line into the CEO, but essentially reported to the board. That created a very insulated legal function.
When Bhavna was describing it, she was using Workday as an example. Workday’s legal function is extremely integrated into the business and is proactive, not reactive.
When we would engage legal, it was after something bad happened.
I feel like the future of the internal legal function—AI is going to enable it. Hopefully, they don’t just reduce headcount and make no promises, because public companies have investors to keep happy.
But my hope is that AI starts taking away some of the grunt work that these legal functions have been bogged down in and allows them to get more integrated into the business and have conversations like the one that you described.
Mark Shelton, who is GC at Barclays and head of the University of Pennsylvania, had a lovely line we would always discuss: “There’s no P in legal.”
It’s cost. I think that’s the very future of legal: making sure that we’re aligning to solving problems with people and not getting caught up in the administrative piece, and giving business meaning and outcomes to people.
Hotpur: That is so important. My first client involved persuading a CFO that overcoming the legal problem was worth spending what we estimated would be $10 million on an AI system in 2011.
As I mentioned, you talk to people about AI today and ask when it arose. According to AI, I hadn’t even been invented at that point.
But it’s all about bringing legal to the table and exposing and bringing to the table the value they can bring. When you merge that with other data and other information, it’s suddenly very, very powerful.
Host: There’s no P in legal, but there sure can be a big L, right?
The same is true with risk management. There’s no P in risk management either. But if that function fails to do its job, my listeners have heard me talk about the four lines of defense.
At Bank of America, we talked about the four lines of defense. The first line of defense is the line of business: the consumer bank, the investment bank, the wealth bank.
Then there’s the second line of defense, which is the risk management function, where AML sits, along with corporate compliance and regulatory relations.
The third line of defense is corporate audit. I also spent several years in corporate audit.
If the first three fail, the fourth line of defense kicks in, and that’s *The Wall Street Journal*, because that’s where you end up.
If the regulators were to come in and write a sanction or a cease and desist—I mean, this has happened at huge financial institutions—you don’t hear about law firms having that level. It’s very rare that you end up in *The Wall Street Journal*.
But you’re starting to hear about it directly related to AI, with lawyers at white-shoe law firms submitting court documents with hallucinated citations.
I kind of feel like we’re more vulnerable now that we’re becoming tech-enabled as an industry. Would you agree?
Hotpur: We are, but I think it’s going to get worse because we’re still solving legal problems.
Actually, there’s a journey. When we start solving big business problems and get it wrong there, there’s even further to fall.
That’s not to say we shouldn’t be heading in that direction. I absolutely think we should be heading in that direction. But we’ve got to really understand how we change, how we operate, the process by which we operate as lawyers, and how we add that extra value.
For me, most of the cases—I can’t claim to have seen all of them, and you’ve probably seen Charleston’s database, for example, of the hallucinations in various court cases.
A lot of them come down to process. They come down to supervision, lack of training, understanding the pressure on someone to deliver something, and frankly, not following their training.
As practicing lawyers, we’ve all had our moment where we’ve not spent the time and researched and done that.
We typically did it in a moment of low stakes, and the partner we were working for and supervising corrected it and got that.
I think too much of the profession is becoming a bit too lazy—being bowled over by the technology, being quite taken by it, and just making assumptions about what it can do and therefore what we don’t need to do, and forgetting that accountability.
It’s actually a topic very close to my heart at the moment. I’m the current co-chair of the Technology and Law Committee at the Law Society of England and Wales, and we’re currently penning the practice note for use of AI by the profession.
We really need to spell it out and educate people around the expectations.
I think we’ve unfortunately done a very poor job as an industry, and the various bodies that are there have done a poor job as well.
It’s very easy to say, “You’re still accountable.” We need to go further than that. Therefore, what do you need to do?
I ran an exercise just a couple of months ago for one of the top 20 law firms here in England and Wales. I’m stunned at what I had to do.
This was a project about helping them with their data strategy. It morphed into having to persuade the lawyers why data is important.
I never thought I’d be in a world where I had to take the SRA solicitor rules and say, “Well, this is what happens if you don’t get your data right.”
You overcharge a client. Guess what the consequences of that are?
You muck up your client money in terms of how you deal with that, and again, you can be struck off as part of it.
Even as someone who’s been in the profession, I was truly stunned by how quick we were to distance ourselves from data that would come to roost in a really poor way if we then overlaid AI through it.
I’ve just worked with a conveyancing firm. They’re a niche conveyancing firm.
They decided not to go with the AI that comes within their practice management system. They decided not to go for one of the big boys with the adverts all over Times Square, et cetera, et cetera.
They’ve decided to build it themselves. And it’s interesting that there’s a blank in their practice management system.
That within, and I think you've got a database background, it will probably get filled with some date, like the first of January 1900 or something like that. You know, catching that later and figuring out why the AI went wrong, and just setting these traps for you going forwards—there's a lot of nasties hidden there.
And, you know, the tragedy for me is we somehow feel it is beneath us to roll up our sleeves and get the data right. You know, there's a sense of entitlement that I've studied so hard—a cluster of what is really long hours—this is my time, why are you getting me to fix the data?
You know, in the stories I'm hearing around, “Well, the partner just gives it to the PA to do.” Even though the data and getting it right is so important. The associate gives it to the PA to do, you know, and creating this culture whereby we don't create that world where the PA can go say, “Well, how am I meant to answer this question?”
Because if you haven't defined the data point, and the data governance isn't there, it's meaningless. Two or three very good lawyers, if they did spend time on it, would give you a different answer.
And that played out. I talked about the bank where I was filling out that form around the rating-downgrade clauses in the securitization hedging. For me, it was really insightful.
I remember very clearly, because I got this list of questions: “What do you mean by that?” And they honestly had no idea. Something as important as, “Well, what do you mean by what's the rating at which this happens?”
Because the document sometimes expresses, “When you hit this rating, then the action occurs,” or, “When you go below it, it occurs.” And they'd not defined that.
So I could straightaway tell you that this bank was going to have inconsistent data, because different people had answered it in different ways. The poor kind of quant downstream, which was a throwback to my days—that led to my journey into the new firm—just assumed there were smart people and they knew what they were doing.
The one thing I will say is, I'm very glad I followed my instinct, because I tried to work with lawyers when I was still on the technology side as a consultant—UPS—and, you know, both internally and also working with industry players and leading law firms, except I struggled, because the answer was, “You're not a lawyer. You don't understand it.”
And my instinct was that I needed to become a lawyer such that they would even just listen to me.
I mean, my wife's the one that paid the price, right? The story is that I proposed and apparently my next words, as soon as she'd executed “yes,” was, “I'm going to be a student and you're going to fund it.”
So my wife was the one that paid the price for it in many ways. But I think that was so important, because even with being able to say that—
The data governance and data quality issues will surface.
And, you know, we're going to be, and we already are being, very quick to blame the tools. And you should never be blaming your tools, right? You've probably got the right or the wrong tool.
And, again, you know, a particular bugbear of mine is people talking about AI and not realizing that it's a suite of tools and ways of working. It's not just about the tool, but the process you work with it and the data to import, et cetera.
Gosh, you know, sometimes I'm saddened, right? Because we think that AI moment has hit, and I think, in itself, we've got a long, long way to go.
You know, and so many aspects of the risks that you highlight. When I look at industries that have mature risk-management practices, again, banks are at the top of that list.
And think about the three—I used to jokingly call them the Wall Street Journal—but let's take the three real lines of defense. You don't have those in law firms.
I've never seen a law firm with a risk-management group of risk professionals that sit and align themselves to a practice area and work in tandem with the practice to ensure that all the compliance boxes are properly checked.
And then, having on top of that, an internal audit team that comes in and evaluates the control environment to ensure that the process has another set of eyes to say, “Okay, I know our policy says that we check our citations and that they're not fabricated. Let me see your controls. What happens if somebody ignores that policy?”
We've seen it play out many, many times now. You need those layers in order to ensure that the control environment fits the risk profile.
And I think the reason that we don't have that is because partner capital would have to fund it. Right? Like, that's a non—there's no P in that, to your point. Right?
And there haven't been a lot of Ls there. There haven't been, until recently, that many—
What was it for the internet? Yeah, see, in-house. So, you know, one bank that I was working with, they had a master agreement—a trading master agreement—in place with the Royal Bank of Scotland.
Now, we have something called close-out netting, and you normally get a legal opinion to say that the close-out netting is enforceable.
Now, these opinions run—for the England and Wales opinion—close to 300 pages. And a large investment bank will maintain about a thousand of these legal opinions and save about a billion dollars of regulatory capital if they can show they have a legal opinion that says that trading agreement is enforceable in terms of its close-out netting provisions.
So this bank said, “Yeah, we've got an opinion,” and they ticked the nettable flag.
I mean, that's the tragedy of it itself, right? The law firm gets away with writing close to 300 pages, chucks it over the fence to the bank, who then has to turn it into a yes or no in their systems.
So that is a live illustration for me of the law firm's saying, “This is where my role stops. And I'm not going to add any more value, right? I'm not going to be a real part of that billion-dollar regulatory-capital saving. I'm only going to be involved in the very legal component of it.”
Now, let me start with what happened here.
It turns out that they did have a legal opinion, but in their systems they tracked it as saying that they had a close-out netting opinion for England and Wales.
And I'll never forget that I had to call up the GC and say, “You've got a problem here, and the regulators are not happy.”
And he said, “Well, what do you mean?”
I said, “Well, your agreements with the Royal Bank of Scotland—the clue's in the name, Scotland. It's got a separate jurisdiction. You should be tying it to the Scottish opinion.”
There was no control framework to do that. And I tell you what it was: it was the lawyers were too lazy to—you thought it beneath themselves to get involved in the data. And the data didn't allow you to pick England or Wales; it didn't allow you to pick Scotland. It hadn't catered.
Now, if we are going to move to a world where we are helping solve the entire business problem, I'm going to have to, because otherwise our clients—be it in-house or law firms—they're going to say, “Well, I can do a lot of this stuff myself using AI.”
They're not going to be perfect. They're going to miss some of the nuances, but the cost saving is part of it, and it's so exciting. And there's no hiding that, right?
You know, these barriers that we've built up—that people have actually not been a lawyer, “You don't get it”—those barriers are gone, right?
You know, maybe all of a sudden—maybe all of a sudden, just wait—that will be available through—
Hang on, right? She is a computer scientist, by the way, so maybe she could have said that.
So, you know, it's going to come because we're going to have to chase the bigger piece of the business problem. And at that point, there's a lot more to gain, but there's a hell of a lot more to fall if we get it wrong.
And then we're going to get the three lines of defense. We're going to get ways of working. And actually, one of the things that I'm very passionate about is there's a lot of talk about smart contracts, et cetera.
A legal opinion is very much at the heart of what a lawyer does. They give their opinion on legal matters, right? And particularly in the US, it's a very established practice. And actually, law firms have committees that talk about the risks and what they will accept and what they won't accept, et cetera, as part of it.
One of the things that I've been looking at for the last few years is smart legal opinions and how we can use that to connect the advice we give as lawyers to be more part of the business, as opposed to having all these intermediaries that have to—
This is a crazy situation on these close-out netting opinions, where these lawyers that do nothing but these close-out netting opinions pass it over to the bank. Some of the biggest US banks have 20 people who just specialize in reviewing those opinions, full-time people.
We're writing the advice in a way that it needs another set of lawyers to even read what it is and turn it into a yes or no before it even gets anywhere near to the business. It's just got to make its way into saying, “Well, what's the regulatory-capital calculation?”
I've been exploring that. As you know, I'm a professor at the University of Surrey. I'm supervising a PhD student who actually retired as a lawyer and was one of these experts on close-out netting. He's just writing up at the moment and trying to look for the possibilities here.
You know, the law changes. How can we create a world where the law changes, the legal opinion gets updated, and it automatically causes some sort of action that means the business does the right thing because of what the legal advice says?
I mean, it's not going to be straightforward. It's going to be fraught with difficulties, but I love your positioning of the control frameworks that you need around there.
But if you get that right, just think of the prize in terms of what we can achieve as part of it, right? In terms of law genuinely accelerating how we empower business, how we empower—you know, it will have wider society benefits for law in many ways.
Yeah, I think that's going to be one pressure that the legal industry is going to feel as the commodity work starts to get eaten away at. You know, there's going to need to be an elevation of use of lawyers' brains, and it's going to force them to be more integrated and more connected to the business.
So, you know, as an entrepreneur for many decades now, I started my journey—I wasn't even old enough to drink when I started my first business. So it was a collection agency, and we used to work with legal partners quite a bit, as you could imagine.
Then I went into the corporate world and then back as an entrepreneur. So I've had lots of exposure to the relationship between client and external counsel, internal client to inside legal functions.
And I can tell you that my experience has always been of an arms-length nature. And I think that I'm not alone in my feelings around that.
So this is a great opportunity. I don't think that it's a lack of curiosity. I do feel that there are dynamics that make the relationship the way it is—somewhat arms-length—but I feel like the barriers between client and law firm are starting to lower as a result of technology and as we start to eat away at the lower tiers of legal work.
There's only one direction to go, and that's up the complexity scale. And we're entering into a world now with AI where we're going to have a much more complex regulatory environment.
I mean, think about just robotics and self-driving cars, and all of the new regulatory frameworks that are going to have to cover that. As the world gets more technical, it gets more complex.
And I think it would be welcomed by lawyers to engage with and understand the business in a deeper way. It's just, I think with law firms specifically, there is a cultural dynamic that needs to be pivoted in terms of how they engage with clients.
I'll go back to Björn again. The reason I asked him to be on the podcast is I saw another podcast he did, and he had a real zinger of a line. He goes, “I have never seen an industry less curious about their clients than law firms,” which I thought was interesting.
And I think it's a perception, because I don't think it's a lack of curiosity. I think it's a cultural dynamic. I don't know—what are your thoughts on what's driving the perception that my legal partners aren't really that curious about how my business operates?
Just wanting to put out fires. I think it's been a function of the business model being so good for so long. In terms of, you know, you bring in super-smart, academically insecure people, put them into this hothouse of being together—a very culture of long hours and demands on billability—that you don't have time to think and fail fast and be innovative in the true sense.
And I think it means that we're always clutching at things and doing them at a supersonic speed.
Another area that I work in is digital assets. So for me, that's a nice mix, right? It's technology and increasingly a lot of law and regulation issues.
I'm actually pleased to say that I've helped write the virtual-asset regulations for the Philippines SEC recently, and Cambodia, and we've got a couple of other jurisdictions on the way that we're assisting with.
And it's been a real struggle. We need lawyers at the table there, because there are things that very much are their purview. But the understanding is so, so shallow.
You know, when I talk about cryptocurrencies, et cetera, I ended up writing the compliance-risk framework for BlockFi. So they were one of the—you know, sort of all good until they took a loan from FTX and never had a good idea.
But, you know, when you talk to people about what actually happens, there's very little understanding and little curiosity that you need. You get statements such as, “Well, it's immutable,” but they're not really. But there's a whole load of assumptions that you've made in order to get there.
And actually, that reminds me—you know, now I'm kind of going down this train of thought—that reminds me of the financial crisis. And I think I said that I worked on these securitization structures.
The rating agencies were giving them AAA ratings, and more AAA ratings, and because somebody did it one day, other people did it.
And if you did ask the question, “But how does it work? Why has it got a triple-A rating?” everyone would just look at you and say, “Don't be stupid.”
And you're almost—you had so much work, you're encouraged not to be curious.
And another very happy memory I have is, you know, one of the big investment banks called up the other partner I worked for and said, “We need the documents done overnight, and we're going to close the entire deal tomorrow.”
And I remember going into the partner's office and saying, “Look, I've cancelled whatever plans I have this evening. I'm very, very happy to do it. Please don't get me wrong, right? But I can honestly tell you, I don't think it's humanly possible to get all of the documents right. It's too complex, and we really can't understand what's going on.”
And I will be forever grateful for that partner taking a view that—well, he listened to me. I'd like to think I earned that respect. And he knew that I was totally serious when I said, “Look, I'll do it.”
You know, that very deal, the law firm that did take it on ended up getting sued for negligence. And I can tell you exactly what it is: we don't create that space because of the business model for people to be curious.
And the technology is moving so quickly. You know, I did a computer science degree at Cambridge. I had the benefit of one of my professors, Michael Lynch. I had Demis Hassabis in the years—a couple of years up from me, et cetera.
You were amongst giants, right? Noam Pricer, as to be, et cetera. John Dalgman, who actually was a professor at Harvard in medicine, then computer scientists, taught me neural networks. He had come up with neural networks to identify virus recognition in the nineties.
It's all about having that. People in technology are struggling to understand, to keep up with the changes that are happening. It's impacting every aspect of our lives.
It's just not possible for lawyers to say it doesn't impact us, because our clients are using technologies. It's impacting us as lawyers.
And it's a really scary time for me in that regard, right? Because you've got to change the model, right?
And I think we're taking it to the extreme, right? Because every partner is saying, well, you know—asking turkeys to vote for Christmas—and say, “Well, you know, it will be the next generation that needs to worry about it. Just let me get across the line and put my feet up and retire.”
And I don't think we're doing enough for the profession and realizing that the pace of change at the moment is just—we just can't afford to do that.
Yeah, so speaking of that and the next generation…
I recorded an episode that should be out ahead of this one with Mike Schmidtberger from Norm Law today. And it was a fascinating conversation.
So Norm Law is what I would call a second-generation AI-native firm. They have the who’s who of former regulators on their advisory board—former chair of the SEC, I mean, just real heavy hitters. Mike was managing partner at [inaudible] for a decade-plus.
They are taking a different model, and they are taking the best of what Big Law brings, which is all of that human—the most talented legal professionals on the planet—and an advisory board with a very risk-focused lens through which they look. Then you’ve got the AI-native platform that doesn’t have all the baggage that a traditional law firm has, like compensation models that are completely out of date when you move away from the billable hour; client engagement models, like we’ve been talking about, that were designed 40 years ago; and capital structures.
They have an MSO. They got $50 million from Blackstone. They’re already at a unicorn. They just raised a round at over a billion dollars. They’ve only been in existence a couple of years.
They don’t have all of that cultural inertia that I was talking about earlier that has to be redirected. And they’re taking swings. It’s not the first generation of AI-native firms that’s just doing contract and data privacy agreement review and NDAs. They’re doing real, substantive legal work—not bet-the-company work, but several rungs higher than what the first generation did on the complexity ladder.
And I’m looking at—and apparently investors are too—the edge that they’re going to have in making change. They’ve got a capital structure that they need funding for. They get it. Law firms don’t have that, right?
In the U.S., anyway, Model Rule 5.4 doesn’t allow for any fee sharing. So you either have to move to Arizona and become an ABS, or you have to leverage an MSO. It takes time, it’s somewhat messy, and for an Am Law firm to do it—and I know several are looking—it was just, I think, a *Financial Times* article that named some firms in the upper echelons of the Am Law that are looking at this model.
But Norm’s already got it, right? They need capital—I’m calling Blackstone. They’re already $50 million in our cap table. And I’m looking at how these big law firms are going to compete.
I know we’re almost out of time, but I’m curious about your thoughts on Big Law’s ability, because they seem to be moving slower than their counterparts on the corporate legal side. I’ve talked about Stephen Crowley from Ford posting on Bloomberg Law, and the title of the article is, “Hey, law firms, you’re not keeping up.”
So we’re behind in the law firm world. We’ve got AI-native structures starting to appear that take the best of what we offer as a Big Law firm, and they’ve got potentially unlimited capital with funders like Blackstone. Then you’ve got these law firms with all this inertia that has to be redirected.
Are we going to be able to make change fast enough before the chickens come home to roost?
It still comes also back to the data, right? I actually think we will find that we talk about the treasure trove of data that a law firm has, but it’s missing the business data that accompanies it in a lot of cases.
So I think that there is a real risk that, if they then suddenly run towards the capital, they’ll all miss the boat if they don’t start moving very quickly.
If I compare it to what some of my large investment bank clients have when their in-house teams start to look at using AI on some of the data they have, and then compare the agreement data with the new European data that they have, with red data, you know, in terms of expressing regulation in some sort of codified form that’s machine-readable, the value is just so much more.
It’s really going to reduce law firms in terms of the value proposition that they provide as part of it.
And I’ve written an article that was published by the Law Society, because if we’re not careful, we’re going to reduce ourselves to being the insurance policy. You go to a lawyer when you want to have someone to sue if it goes wrong. That will be our function.
There’s so much—my journey’s shown me that there’s so much value lawyers can bring. I think the U.S. actually does a lot better at that, right? You look at where people who have worked in Big Law and some of the roles they take on the business side afterward. I’m not sure we’ve got that same culture on the other side of the pond.
But it’s going to be a challenge. And I’m fearful. I feel I have a love for the profession, and I feel that it may just take one or two incidents—high-profile incidents—and it might just force the issue.
But it’s not, I don’t think, just going to happen because everyone thinks it’s the right thing to do.
Yeah. I mean, the good news is we’re not out of time yet. We will be if we don’t start the work in earnest now.
There’s still a lot of experimentation happening, and hey, I get that. I hear people ask questions about ROI.
I talk about the three waves of this transformation. The first wave is the co-pilot era, which is with a lower KC—you know, the Harveys, Legoras, the world.
The second wave is where we re-engineer the process and unbundle the legal work, and enable it to scale with technology.
Then the third wave is the autopilot, or the agentic era.
We are still figuring out wave one, and these AI-native firms are starting in wave two. They’re not using workflows from 1995 that they bolt a co-pilot onto. They’re starting in wave two, which is light-years ahead of where we are.
So I’m hopeful that candid conversations like this inspire a sense of urgency. I don’t think panic is the right path, but a sense of urgency—moving in earnest, making the investments, and stopping the pearl-clutching in the industry that says, “Hey, we’re making record profits.”
What worked yesterday won’t work tomorrow.
Well, investing in people, right, and investing in our future generations, and realizing that this isn’t about replacing the junior work, because we’ll create another problem as part of it, right?
The reason we sponsor the PhD is because it’s an investment in the future, right? And that’s got to be the way you look at it.
I guess I’m opening up a whole can of worms there in terms of what do we do? Do the junior members of our profession equally have the maturity to realize that it’s so crucial that they don’t take the shortcut and use AI and kind of remove their ability to learn?
I have a lot of fun explaining to people how neural networks actually operate, and the fact that it’s all about feeding it more data, trying things, getting things wrong, trying again, et cetera. And the moment the junior associate, the trainee, gives that to the AI to do is actually when they suffer.
But again, it comes back to the business model, right? The pressure is such that you’re forced, in many ways, to just look for the fastest way to get it out for the client demands.
One thing we haven’t touched on is, you know, the client’s always right.
Maybe this is just what it is. Maybe clients have got to play a part in that. I know clients are equally bashing law firms to say, “Look, if the AI can do it fast, and therefore you really need to charge me that. We want you to use AI as part of it,” et cetera.
But maybe they need to play a part in that as well, right? And be a bit more mature and foresighted in terms of ultimately where we’re trying to get to.
Yeah, I think the challenge with that—I agree with you—is that there’s so much pressure coming from shareholders, going to boards, down to management, to deliver savings as a result of this.
I think it’s going to impair clients from thinking long term, because if you’re at a publicly traded company, you’re beholden to shareholder demands, and they don’t really care about the profession.
But what you’re saying there buys into the view that AI—what will it do to society? It will just reduce jobs, it will reduce cost. It’s not really buying into that view that, no, it’s value creation.
As with all other big advances, it actually changes the very nature of what we can achieve and work, right? That’s implicit in the statement that you’ve made there.
But you’re right, it’s those short-term pressures that we’ve put through shareholders or through corporate investor demand.
But, for the brief view, who tried the other way, as you say, the AI natives—and maybe they’ll just put enough pressure to mean that people realize, you know, they’ll—
—than you’ve got bus drivers, if they’re not careful.
Indeed.
Indeed. I use Netflix and Blockbuster quite a bit. There are lots of lessons, I think, we can learn from how that story played out.
But this has been a fantastic conversation. I knew it would be. Thank you so much for staying up late on your end.
How do people find out more about D2 and yourself?
D2LegalTek.com. We do a lot of work in the derivatives space. There are lots of [inaudible] out there, and hopefully it can help people think through some of the challenges and opportunities that we have.
We’d love to have more conversations with the UTA and others on it. So please get in touch, anyone.
That sounds great. All right, thanks again, and have a good night, Akbar.
Thank you.
All right.
Thanks for listening to *Legal Innovations Spotlight*. If you found value in this chat, hit the subscribe button to be notified when we release new episodes.
We’d also really appreciate it if you could take a moment to rate us and leave us a review wherever you’re listening right now. Your feedback helps us provide you with top-notch content.