AI's ability to automate is transforming traditional workflows
and transactions, but in this fast-paced era, it can be difficult
to predict the full extent of AI's influence on complex tasks
such as due diligence and dealmaking. In this podcast, our New York
tax partner Scott Semer is joined by Nick Kato, founder of Leo
Berwick and co-founder of Elbe, to explore the impact of AI on
professional services, and what the future holds for the role of
dealmakers, lawyers and others in the professional services
sector.
Transcript
Introductory Speaker: Hello. Hello and welcome to the Torys Live podcast. We are delighted to have our special guest today, Nick Kato. He is the founder of Leo Berwick and the co-founder of Elbe. Nick is joining us today to discuss artificial intelligence. So Nick, thank you for joining us.
00:28 - 00:30
Nick Kato (NK): Thank you for having me. I'm excited to be here.
00:30 - 00:40
Introductory Speaker: Yeah. Well, I'm going to turn you over now to our host, our esteemed tax partner from New York, Scott. Semer. Scott, I will let you take it away.
00:41 - 00:58
Scott Semer (SS): Great. Thanks very much. Thanks a lot for joining us, Nick. Hoping to have a really exciting discussion about artificial intelligence. Just to open up the discussion, I'm curious if you could ... if someone asks you to describe AI for dummies, what is it? What do we mean when we say artificial intelligence?
00:58 - 01:22
NK: Yeah. I mean, there's a lot of different things around AI, you know, and, you know when we look at ChatGPT, for example, right, you know, it's really like an advanced form of Google, you know, which I think is really cool. It can create things for you, take to another level, you know, it's more humanistic. You know, AI can also think a little deeper, a little harder than your typical, you know, Google searches or things like that. And so, you know, AI comes in a lot of different forms, but, you know, I would say the most common form that I see it being used in, in what we're doing is when in a machine learning aspect of AI—kind of training models to, you know, think like humans or even better than humans, quite frankly—I think that's really the future of AI and professional services.
01:42 - 02:03
SS: It's interesting. As someone who has founded a professional services company and also founded an AI company, I am curious to your thoughts on, you know, what it is, what should law firms and professional service firms that let's say, don't have, you know, aren't as big as an EUI and don't have hundreds of millions of dollars that throw at AI program, what do you think they shouldn't be doing with respect to AI?
02:07 - 04:48
NK: Yeah, it's a good question. And, you know, I actually think that professional service firms at it's at their core are not the best at technology, quite frankly. And there's been an opportunity even over the last 20 years to utilize it in, in our fields and it hasn't been done. And, you know, even without AI, like basic automation of tasks, I think it's something that isheavily overlooked, you know. There was a lot of, a lot of people think that, you know, practicing law, you know, isn't ... can't be automated, right? Which I agree, at its core, it can't be. I mean, but there are a lot of aspects of running a project, you know, as a lawyer, that can be and if you're doing a due diligence assignment, for example, or from contract review or negotiating an SPA, you know, I think there are a lot of things there, you know, that can be automated. And, you know, even before you get to the AI front, just through, you know, technology and processes and whatnot, you know. When you get to the AI aspect of things, you know, obviously there are firms like Kira, who got bought by Litera a couple years ago, which is used in a lot of law firms currently for contract review and things like that. You know, I see that technology expanding and, and getting, you know, bigger and better and broader, you know, across the other services that, you know, law firms provide. You know, there are firms out there like Blue J (Legal), for example, and some other ones that are trying to in the tax research and kind of using AI to help facilitate tax research. They've been working on that for a while and I've seen some progress there as well that could definitely be an avenue as well. But you know, I don't think that a lot of firms need to spend hundreds of millions of dollars to develop their own, you know, machine learning, you know, programs, deep AI, you know, projects and whatnot.
I think kind of basic automation with technology can go a long way. And it took me a long time to figure that out, quite frankly. Because when I first was told that we should be thinking about technology in our business by a mentor of mine, maybe two-three years ago, I kind of said, that's not possible. You know, you can't really use tech in what we do. And as I kind of thought about it, more and more, I've realized that there are just so many ways to use technology. Not even I just kind of the automation of tasks to make our lives easier and our clients lives easier. And I think that I think should be an area of focus for, all firms, quite frankly.
04:49 - 05:04
SS: What do you think is like one of the most exciting of those areas that you, you think about that will be helping you in your practice and helping law firms like where is the area, sort of the low hanging fruit that's really ripe for, the benefits of AI?
05:04 - 07:18
NK: Yeah. For sure. So I can tell you what we're doing at our company Elbe, for example, and I think, you know, that can give you a basis for how it can be used elsewhere. But, you know, what we're doing at Elbe is we're automating the due diligence process, right. And so legal, financial, tax, HR, IT, all of that. And to think about, I know this is for Torys, think about legal due diligence. Right? And the questions that are asked are in that process. A lot of that, you know, are the same questions are asked on every deal. And sure, a different industry of different questions, right, but that can all be kind of accounted for, right. and you can kind of come up with those list of questions and then you can say, well, you know, it's the key that when diligence, as we all know, is the follow up questions. But you can kind of narrow down to the follow up questions are also, if you think it all through this, you can create kind of complex decision trees. And then they kind of almost answer – ask and answer -could answer many, many of the diligence questions that you already otherwise ask. Right. And so that's what we're doing at Elbe right now.
And then as part of our software, we're also kind of creating an updated information request list where, you know, as the questions are answered, the information request list is updated. And then we have a captive data room as well, which will grab files from, you know, someone's hard drive if they want to satisfy the information requests lists that are already up there, scan the documents and then pre-populate answers to the questions in the diligence questionnaires, and also create new questions, you know, based off of a scanning of those documents as well.
And if you think about that, you know, that's a lot of what diligence is. And it would, you know, really expedite things if you, for example, if you're a buyer of ... if you're a private equity firm funded by a company and you're to send them our software, the target company, they could start that diligence process immediately, right? And they would see, you know, a lot and you know the process and make things a lot more efficient. Right. And so there's a lot of things on the diligence side that I think can get done through automation in AI and that's kind of where we're starting at our company.
07:20 - 07:43
SS: And so what's the what's the main advantage? So if you're a private equity firm and you're a general counsel or someone in the legal group there, what is the big advantage to you of AI? Is a cost savings because things they use have to pay someone like Leo Berwick or Torys to do can now be done much cheaper. Is it speed? Is it more thorough? Is it a combination of all those things?
07:44 - 08:56
NK: So I think, you know, from, from my perspective, there could be a cost savings element to things for sure. But it's speed, I think. And so my good the context of a deal, for example, right, a lot of times, you know, private equity firms want to hold off on spend with advisors until they have more certainty around the transaction, right, and then they have this mad dash for all their diligence done in a month or two, right? And as we all know, that drags on. Things don't get done, right, deal drags on and, you know, time can kill deals, which isn't great. And so I think efficiency, speed and if making the process much smoother, you know, is a big element of how technology can kind of be used in, you know, transactions to, you know, just make the process better and faster, and more efficient. And, you know, with that may come savings in fees as well. But I actually think that, what's going to happen is that firms are going to utilize this technology, there might be some efficiencies gained through it, right, their fees may be lower potentially in the future, but margin will be higher.
08:57 - 09:23
SS: And what do you think about if you're a PE firm? So you think PE firms will internalize this AI so things that they used to outsource they can now do internally because they'll have AI? Or do you think it'll still be the case that they'll, they'll hire an outside advisor. But the outside advisor will be a much more efficient. They'll looking for an outside advisor that incorporates AI into their process and procedures. Or it'll be kind of some combination of that?
09:24 - 11:38
NK: I think it's a combination. You know, like I don't, believe, at least in the next five to 10 years—in 10 years, call it—that AI is going to replace humans, right. You hear Elon talking about, you know, AI and the future and all that. And you know, if he's right then, you know, potentially it could. But I think in the, you know, next decade or so I don't see that happening.
I think that humans are going to be necessary no matter what in, in the next decade or so and maybe thereafter, also, really, it depends on how crazy the stuff advances and spools up, right. But I do think that you need a human to kind of just look at all the factors, you know, like when you're looking on a tax project, you know, Scott, for example, you know, it being able to connect the dots in kind of abstract ways, right? And it's it would be hard to train a machine to do that, you know, initially. Right. Now, maybe over years and years and years it's possible, but it requires training. It's not like you just create AI that does all the stuff magically for you. It requires training the models to do so. And you know, that requires time and, you know, a lot of, expertise. And I, it's funny, I actually think that the biggest obstacle in developing technology, for you know, a tax advisor or a lawyer or, you know, an accountant and so on and so forth, is having the engineers and the service providers work together to make things happen. Like, we've been working on our product for about a year and a half now and that's been one of the bigger obstacles, quite frankly, is trying to get the, the, the subject matter experts to learn to kind of train the models with the engineers, right. And it, one, takes a lot of time, but it takes a certain mindset which a lot of advisors don't have, quite frankly.
And, you know, I think that to me would be the greatest obstacle going forward to make, you know, AI as prevalent as it needs to be, is you need to get getting lots of time dedicated from lots of experts to sit down there and train these models to make it all happen. And it's not easy.
11:40 - 12:18
SS: Yeah. I mean, that brings up an interesting question, which is if you're training your replacement, right, what's the sort of incentive to do that? And then that kind of goes into the question of how do law firms and service firms manage their pricing models, which have traditionally been kind of a cost play, cost plus base model based on hours worked, and you charge kind of a, price for hours worked, which is kind of a proxy for the work you do. If you're training a machine that can kind of do that, you know, almost instantaneously, and therefore it's going to take you much less hours, how do you capture all of that effort that went into training in terms of your pricing models?
12:19 - 13:26
NK: Yeah, I know it's a great question, honestly. You know, and I think that, in my mind, you know, there will be, it's a math question, quite frankly, right. It's like, what will the market bear, right? And what are companies willing to bear as far as the expense to make this all happen? And so, you know, I think that in the coming years, there's going to be a company or two that emerge as leaders, you know, in, in these sectors, right. And they'll kind of monopolize a lot of this technology and whatnot and it'll just be licensed out to people, quite frankly. And I think that, you know, that's where this will go. I think if you go to model of like, each firm tries to create their own product and then, you know, monetize that, you know, in the market, I think that is a much less efficient process and probably won't work quite as well in my mind.
And so, but if a firm want to go do that, then, yeah, it would have to try to find a way to, you know, recover that investment over a period of time and be comfortable making that investment as well.
13:27 - 13:38
SS: Yes. Are we, are we going to see, you know, law firms, where law firms are paying the licensing fee and they have fewer lawyers and each lawyer is charging, you know, instead of $1,000 an hour or $10,000 an hour.
13:40 -14:04
NK: I think you can maybe see it going that route or a way to maybe amass some of that hourly fee is flat fee arrangements on projects as well potentially. But I could see yeah rates going up you know for sure at firms for expertise. And that goes back to my point earlier, which is I can see fees and professional services go getting lower over time, but margin increasing over time.
14:05 - 14:33
SS: Yeah. Yeah. Fascinating. That brings me to another question, which is, for, you know, a lot of people listening to this podcast, a lot of them have children who are going to college. Maybe they're considering going to law school. A professional services firms, you know, should should all of us advise our children to become plumbers and, electrical engineers rather than going to fields like law and accounting where, they're going to be ultimately replaced by AI?
14:34 - 16:11
NK: I don't think so. I mean, I just don't think AI is going to replace, you know, our profession. I think it's going to bring in a lot of efficiency though, quite frankly, right. And so I think there'll be less of a need for talent in the future, quite frankly. And I see that happening for sure in the market. And so I think that, you know, we should if our kids want to go, you know, practice law, they should go do that. And there'll still be ample jobs in the future in that space. But I do think there'll be less than there currently are. And, you know, I think that AI and automation, a lot of that is going to, quite frankly, you know, require less junior resources. Like I think it's going to it's not going to replace the senior partner on the file, right, but I think it could replace some of the junior associates, for example. But you'll still need junior associates to do some of the work, right, and I think that to me is a really interesting issue because the question becomes how do you train, you know, lawyers fresh out of school, right, if you have a lot of this AI and this kind of automation, you know, type stuff going around? Because, you know, they're not going to learn the same way that you and I learnt, Scott, kind of coming up, right they're going to learn in a very different way. And I think that's something that should probably be studied and be thought about a little harder because I think that's coming, right? You know, when you have these new graduates coming out, how are we training them? What does that look like? Because I think it will be different.
16:12 - 16:52
SS: Yeah, that's very interesting, right. You see that a lot of other industries where there's a much deeper learning curve and you either kind of hit the top pretty quickly or there is just not a lot of room for you in there, as opposed to the traditional law firm and accounting firm model.
Curious, since you do a lot of work in infrastructure investing, you know, a lot of your clients invest in infrastructure, one of the inputs for AI, right, is it's pretty electricity intensive. Curious if you see kind of another aspect of AI is this increasing demand for energy and whether there's, things people really should be focusing on in terms of investing opportunities there?
16:53 - 18:09
NK: Absolutely, right. You know, you, you know, you're seeing all these data centers pop up everywhere. And, you know, how do you, you know, power these things, right? They require a lot of energy. And then, you know, AI in general, too, requires a lot of energy, you know, to do things. And so, yes, you know, generation is going to be a huge thing, you know, how do we how do we generate the electricity that we need to, you know, support, you know, AI in the future? And we're seeing a ton of investors focus on this, you know, especially in the energy transition. So like renewable energy, introducing a ton of investment in renewable energy right now, battery storage, you know, and what we're starting to see a trend of as well as transmission because I think it's one thing it's great to generate all this power but as things scale and other, you know, the footprint grows, it's going to be necessary be able to transmit this electricity to locations where there will be a need. And so we're seeing a big pickup there as well. And so you're right, Scott, even though we're talking about tech here, it does spill over into infrastructure quite a bit because there's a big demand for energy. And you're seeing it with our clients. They're investing heavily in that sector. Including the adjacent areas like transmission.
18:11 - 18: 20
SS: It's interesting. You know, curious, whether that's ultimately going to be kind of what determines the rate of growth for AI is just is there enough power generated to keep up with the demand?
18:21 - 18:39
NK: Yeah, absolutely. And you know, I know we're all working towards energy transition and I think it's an amazing thing, and I think I've seen amazing progress there you know as well. But it does raise the question, you know, is renewable energy going to be enough to support all this. Are we going to have to look to kind of more traditional forms to supplement it?
18:40 - 19:40
SS: If you had to make a prediction, let's say five to 10 years from now, what do you think are going to be the most successful law firms in terms of their ability to use AI? Do you think there'll be just sort of fewer law firms, and we'll see a wave of the mergers that we had, you know, before, because there's such a need for, pretty intensive investment, in whether it's licensing or what it is, in AI and the only way the model works is if you just have a larger firm that where that cost is kind of can be spread around the more clients? Or do you see more specialized firms where they have a particular expertise, where you really need the expertise of the senior partners? And not just law firms, but accounting firms, you know. But if you had to pick, what do you think the world is going to look like in 5 to 10 years in terms of how it's different, in terms of what law firms are successful and what accounting firms are successful, what are the predictions you to offer at this point?
19:41 - 21:40
NK: Yeah, it's a good, it's a really interesting question. You know, I don't think AI is going to cause, you know, a lot of mergers and all of that. Again, because I don't think, you know, one firm is going to go out there and develop this thing, you know. I mean, firms are going to develop this stuff on their own, right. Like, you know, I'm doing it right now, quite frankly, you know, in my other business Elbe. It's, it's very challenging. It's not easy. And it, it's a big distraction for, quite frankly, the billable people, right. I think that's why you haven't seen a lot of even automation, quite frankly, in professional services, let alone AI. It's because of the time it takes, you know, the actual professionals who are billing to go do it, right.
And so what I see happening is I see a few players stepping up in the coming years and, you know, really driving this and making this a focus and then licensing that technology to other firms. And like I said, I think some of the firms' headcount could shrink potentially from where it is today could continue to grow with as the economy grows and whatnot but I think the leverage model could change a bit. And, I think that's where it goes, at least in the next kind of 10 to 20 years. Like 50 years from now, like, who knows, right? You know, maybe, you know, robots around the world, right, but, I think over the next 10-20 years and I see, kind of ... I don't see firms going out of business because of AI. I don't see massive mergers occurring because of that either. Yeah, I just think that the way firms operate is going to change. Their pricing models are going to change. The way they staff engagements are going to change. Where they hire and train are going to change. And I think in my opinion, that's probably the one thing that, I were a professional service firm, I'd be focused on now, is, you know, like, how is our business model going to change over the next five or 10 years as some of this technology rolls out?
21:41 - 22:23
SS: It's a pretty interesting question, right? Especially for an industry that doesn't like change and as you said, likes to spend their time on kind of billable matters and thinking about legal and accounting and sort of substantive issues and not thinking about necessarily a lot of these business issues. So it's been a it's been a fascinating discussion. I want to end it by asking you what, as both a, someone who's trying to create a, you know, AI product and one who potentially uses it and one who potentially could have AI as a competitor, if you will, what makes you most excited about AI in the next, you know, two to three years? Let's use a shorter time period. Like what are you most looking forward to about it?
22:24 - 23:24
NK: Yeah, I mean, I could speak to it from like the Leo Berwick perspective. I think AI, automation technology in general is kind of just make, you know, the, you know, the M&A process a lot better, quite frankly. It's very archaic right now. You know, you think about the last technological development in M&A, the data room, right, was the last one, and so when you look at the future you know, I think in some of the tools that we're building right now that support M&A, I think it's going to be great. I think it's going to make deals a lot easier to get done, and sellers are going to be more prepared, you know, buyers to spend time focusing on, you know, more deal issues, less diligence and information gathering. I think it will just streamline, a lot of things and allow, you know, folks that are more junior maybe to, to learn other skills that are less menial, quite frankly, than they're currently doing.
23:25 - 23:40
SS: Interesting. And do you think that will lead to maybe you know, increased deal activity because it actually be easier to complete deals. And so there may actually be, you know, a lot more opportunity to do some M&A activity that before you just couldn't get the information you need, or it's just too much of a roadblock.
23:41 - 24:55
NK: Yeah, 100%. You know, like, you know, one of our goals with our technology platform, is to have it be a used on every deal going forward. And, you know, if you think about it like if you're a seller or a business, right, and you're getting ready to sell your company, you know, you go, maybe, maybe you're big enough that you go out and hire an investment banker to help you deal with some of the commercial things, some of the financial aspects of things. But, you know, when was the last time, you know, someone from Torys went in and did, you know, legal diligence, you know, on a company and it was just easy, right? Like they had all their files uploaded. They had answers to all the questions already. It was seamless, right. Like the answer is never, probably, right. And I think that, you know, our technology is going to help change that. Yeah, we're going to help sellers get ready, you know, by, you know, having them having them answer 60, 70% of the questions that would already be asked by a potential buyer. Provide 95% of documents, you know, that have already been, you know, you know, that would be asked by a buyer. And it will allow them to get in front of these issues if there are issues identified and be better prepared and or remediated things before they come to market, so it's just a very smooth, seamless process.
24:56 - 25:13
SS: Yeah. I mean, it sounds super exciting. I mean, it's great to think about some of the exciting and positive developments. Really fascinating discussion, Nick. Really appreciate you joining us today and look forward to seeing, what types of AI you guys are able to come up with.
25:14 - 25:16
NK: Yeah. Really excited next time Scott. Appreciate it. Thank you.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.