Data Science Leaders | Episode 32 | 30:06 | January 4, 2022
Legal Analytics: Winning Business, Winning Cases, and Winning Over Your General Counsel
Get new episodes in your inbox
powered by Sounder
Legal work may not be an obvious application of data science to many advanced analytics leaders. But that should change.
In this episode, Peter Geovanes, Head of Data Strategy, AI & Analytics at Winston & Strawn, breaks down the nuts and bolts of legal analytics and how it’s revolutionizing the way law firms win new business—and cases. Plus, he shares insight on the types of legal challenges data science can help address inside any organization.
Welcome to another episode of the Data Science Leaders podcast. Our guest today is Peter Geovanes. Peter, how are you doing today?
Hi Dave. Doing great, thank you. Thanks for having me.
Great. Peter is the Head of Data Strategy, AI, and Analytics at Winston & Strawn. You hear that and you immediately think that it sounds like a law firm. Well, it is! Today's agenda is actually going to be focused on legal analytics. Peter, what on earth is legal analytics?
The best way to answer that, Dave, is to give you a little bit of my background. I started my career as a Naval officer. I did eight years of active duty then went back to school for my MBA. Let's call that chapter one.
Then I embarked on a 20-year consulting career, working at firms like Accenture, PwC, Alvarez & Marsal, and AlixPartners. I want to call out one five year stretch, which was 2000 to 2005, when I worked at SPSS. Any of our data scientists or data people out there may know: SPSS and SAS are probably credited with creating this whole wave that we now call data analytics or data science.
2000 to 2005 was, I hate to say, 20 years ago, when I really got a PhD level education. It was funny. Back then, we'd kind of elbow each other. This is 2001, saying, "I think this is the year this data analytics thing is going to take off." Who knew? What did we know?
Let me now go toward the end of that 20-year span. I guess I had a midlife crisis and decided I don't have enough stress in my life. I went back to law school full time while I continued to work full time. In probably my second year of law school, I had this epiphany.
At SPSS, and subsequently at other consulting firms, I'd been using data and analytics, specifically predictive analytics, to help Fortune 1000 companies solve business problems. I thought to myself, "I've never seen anybody try to incorporate legal data." It was public domain. The data was available.
Was it possible to predict, dare I say, duration, cost, and maybe even outcome of litigation at the very onset, by using some type of modeling technique? I had a little bit of success. I used a survival model, hazard Cox. It seemed to give us some insights into that.
Right around the same time, transitioning to chapter three which is the law firm, Winston & Strawn, I guess they had their own ‘aha’ moment and thought, "You know, it's not going to be too far in the future where data and analytics would maybe augment an attorney's knowledge, intuition, experience. And we could use that for the benefit of our clients to give them a competitive advantage."
Serendipity, we'll call it, but the job req came to me. I will say now, 5+ years later, it was completely wrong. I knew where they wanted to go. I helped them rewrite the job req and I'm thankful to say I got the job.
Got it. When you went into law school, were you going there intending to eventually become a lawyer? Or were you going there thinking about your foresight and background in analytics, thinking that the legal profession could use some analytics?
In consulting, as you get more senior, you start selling work to clients. I found that I was constantly in front of the general counsel, trying to defend the statement of work that I wrote. My immediate thought was, "I don't want to keep going into these so-called battles or discussions unarmed. Even if I can go through one year of law school, I know contracts is one of the courses. I would love to get that knowledge."
Law, as you would probably imagine, is very analytical as well. It became my passion. I really enjoyed it. When the possibility started coming to me of combining these passions, data and analytics, with my newfound appreciation of the law, there were no jobs out there.
I wondered if there was an opportunity to help, whether it was the law firm or perhaps even as a corporation, helping the general counsel. So, that was it. It was that simple of an idea. Thankfully, I was able to take my experience as a JD/MBA data scientist, and maybe help create a new little discipline for us as data scientists.
I think there are a couple of angles to this. It sounds like it’s about helping out. If you're a data science leader today, you're working at a large corporation that has their own legal team and general counsel. You’re helping that team out but you're working for a law firm, right?
There's a sales side to being a lawyer. You have to get clients. I imagine you're helping there. There's also helping Winston & Strawn actually win cases. That would mean helping them better understand the judges that they're going to be in front of, tendencies, and that sort of thing.
What piques my interest is how analytics plays a role in helping win cases. I think we'll eventually get to some of the other aspects as well. Maybe you could talk a little bit about what you've done there.
You're absolutely right. When I think of my role at the firm, I think in two different sections. The business side of law, which is, as you alluded to, marketing, business development, helping the HR team recruit. Those are all fundamental things. No matter what business we're in, those are critical and we know data and analytics can play a huge part in giving us a competitive advantage.
The other aspect that you alluded to is more on the practice side of law. What can we do? One of the first projects that I worked on at the firm was something we called early case assessment. There are a lot of great third party legal vendors that give you information on judges, the opposing counsel, how many times the judge has ruled for the plaintiff or the defendant, how that judge compares to other judges in the same court or district, etc.
I almost think of it as a baseball card. If you turn over a baseball card, you have the statistics of how that player has done. That's kind of what we're doing with our early case assessment. We're trying to give our attorneys collective litigation insights right off the bat. So they know, for example, the opposing counsel that we're going to, they always settle, but they always settle at the 11th hour. That's great to know as an attorney because now you can also brief your client—the CFO or general counsel—this is their typical trend.
Maybe a more extreme example would be the judge. It's like, "Listen, the judge typically rules for the plaintiff here. However, in the instances that they have ruled for the defendant—and I should say our law firm, Winston & Strawn, is typically on the defendant's side—here's the arguments that seem to resonate."
We basically mined all of what’s available in the public domain. We were able to find this judge. We were able to find all the motions that went in front of this judge. We’ve found patterns. If you cite this case or make these arguments, it seems to be, that may win the judge's favor. These are very simple examples. That's one aspect of, I'll call it the practice side of law, where we could give analytics an opportunity to shine.
Hopefully those in the audience will not be surprised to hear that judges have bias themselves, with their own tendencies. We, as practitioners, can look through this data and uncover some of those tendencies to better help your law firm win cases.
I have one other example, if I may share with you. This is software. Some of the applications were home grown. Other applications, now, are products that could be bought off the shelf. We invested about a year and a half ago in an AI. I'm going to call it a due diligence tool.
Typically, for a merger and acquisition, Company A wants to acquire Company B. The first thing that'll happen is the investment banker and the attorneys will get together. They'll, say, collect in a deal room work, for our sake as data scientists basically FTP to a secure site, all the executory contracts. What we want is for the attorneys to go through that and pull out certain data elements. It might be the names, the dates that were executed, the dollar amounts, automatic renewal status—I'm sure there are 20 different fields that they're interested in in this 50-page contract.
In the past, well, you'd get an army of young associates, typically first, second, or third year, looking through 10,000 contracts and maybe collaborating on one spreadsheet to read through these, find the terms, put them in the spreadsheet. Very long. These are expensive resources. Very costly to the client.
With this new software, called Kira, we use AI. It’s specifically NLP and it has already been pre-trained. It knows if this is a commercial contract. The attorney can basically click on the left hand side of the screen and say, "I want parties, dates, area of law, etc." Whatever they want.
The AI NLP will basically read through all these contracts that have been uploaded and pre-fill that spreadsheet for you. If it misses something, that's where we're going to go look for whether something looks odd in the spreadsheet. We can quickly answer which of these contracts are going to expire in the next 60, 90, 180 days. We can also quickly sort them by dollar amount.
What would've taken 1,000 man hours, I can do in one day. It's huge. What's interesting at a law firm is that we bill time and materials, but what's happening is clients are getting more sophisticated. You could fix a bid from, say, 250,000 to 100,000. By bringing technology into this, you can actually increase your margins, making it a win-win for the client. Hopefully, as a law firm, you will also be able to do more and add value, add worth, rather than just cross-eyed looking at the contracts.
If you're thinking more about your billable hours not going down, you're just opening yourself up to competitors.
Law firms compete just like every other company out there. If you don't innovate and come up with better solutions that are faster and potentially cheaper, then you're just going to get undercut, for sure. What's interesting is that you were talking about contract analytics. Dealing with contracts is the bane of many companies' existence. This is not just law firms, but also companies dealing with order forms from a vendor's standpoint and vice versa, on the customer side as well. I imagine that this type of NLP is applicable, not just in the legal world, but beyond that as well.
Yeah, great point. I think that's what we're seeing, too. As I mentioned, clients are getting more sophisticated. They're bringing these tools in-house, whether it's to manage their vendor contracts or whatever type of purchasing agreements. They’re doing this to be able to quickly build a database, extract those elements, have it search within it, to proactively know if something's going to renew in the next 30, 60, 90 days, and bring that to the procurement team.
You might think of it only for a law firm, but I think, to your point, Dave, that it's a much broader market. There are a lot of applications. To the people listening, my experience has been that the general counsel's office, from a BI perspective or data analytics perspective, is typically forgotten. The general counsel has a lot of money and needs. My suggestion would be to endear yourself to them. There are a lot of simple use cases, whether it be creating a dashboard for them, that will absolutely endear you and you'll make a huge difference in your organization.
That makes a lot of sense. The type of work that you're doing there could be very easy to do, but it's never a bad thing to win over your general counsel and have them on your side.
Let's switch over and talk a little bit more on the business side. If you were working specifically for a law firm, working with some of the partners, what might be some of their needs?
As you become a more senior attorney, ultimately you want to make partner. One of the qualifications is your ability to bring business to the firm. There’s no quota per se, but you have to have a book of business. You have to have clients, be well respected in the industry. It's a tough job. Business development is something that has to happen all the time. Attorneys are busy people. They typically aren't working 40 hours a week. It's more like 50 or 60, but it has to be done.
I was at the firm for maybe two or three weeks. I happened to go out to lunch with my friend, Amy, who was a director in the business development department. I'm pretty naive to this so I just asked the question, knowing I may sound silly, "Hey, Amy, help me understand. How does a large law firm like ours conduct business development, and what does your team do to assist them?"
One of the examples she shared, she said, "We subscribe to five different news services. So, every time a lawsuit is filed, whether it be in the state or federal court, we get an email alert and it's going to have some basic information in there. It's going to talk about the nature of the suit. It's going to talk about the venue. Is it federal or district court? State or federal? Who the opposing counsel is, who the plaintiff is. Perhaps the judge has already been assigned."
So I said, "Interesting. So, what happens next?"
She said, "Well, we have a central email box that we send them all to. The next morning I have a team of analysts basically go through these leads, looking for the needle in a haystack."
I said, "Well, how many do you typically get?"
She said, "You know, it varies. But let's say, on average, about 1,000 of these come in a day."
"Okay. Well, how many are typically hot leads?"
"Pretty low. Maybe 5-10."
Okay. So pretty low. I said, "How much time does it take?"
She said, "Well, typically I get two people working anywhere from 8-16 FTE hours a day to get this out, and then we turn it around to the attorneys, hopefully, by the end of the day."
"Interesting." I put my data scientist hat on and I'm thinking, "Okay, by chance, do you happen to keep the ones that you've accepted and rejected—maybe six months worth?"
She said, "Yes, I do."
"This could be automated. This trains the neural network." Right?
There's your training set.
We could basically use the ones that you've done in the past and imply what the heuristics or business rules are that you apply. For example, we do B2B law. So, if someone has a slip-and-fall at the local Walmart, that's not us, nor are the auto accidents. Those will automatically be discarded, but right now, somebody's reading through those.
That's exactly what I did for her. I built this solution which takes the emails and unstructured data, parses them and puts them into a database so it's structured. I do some modeling on top of it. I score everything. I'm able to tell what are the leads and not.
What I didn't think of, Dave, are the really interesting consequences here. Number one is I freed up those two FTEs to do more value-add work. That's pretty good.
Secondly, instead of having that lag of up to 12 hours, we're done instantaneously. We’re potentially the first in the queue to call that general counsel saying, "Hey, I heard about this. I think we can help you." That's pretty good.
The third part, though, is the one that really got me excited and it completely escaped me. I'm parsing every email that comes through but, in essence, I've built a database of every lawsuit in the United States in real time.
I use a technique called augmented analytics. The way I like to think of it is as a virtual assistant. It's constantly looking at the data, trying to find trends, correlations, and outliers then it presents it to the business development team, as much as a physical analyst would.
Let's say that the data shows some litigation going on, perhaps at Ford, General Motors, I don't know, Nissan and Toyota. Maybe you want to do work at Tesla. The systems alerted you that something was going on. Maybe it's with airbags and new litigation. We could pass that information onto the partner that has a relationship and they could call up, perhaps, that general counsel, Mr. or Mrs. General Counsel at Tesla, and say, "Listen, I don't know if this legal problem has hit you yet, but we're seeing this across the industry. More importantly, I'd love to have a conversation with you to talk about what we can do proactively to get ahead of this."
Dave, that changes everything. The law firm is now basically out there as a fiduciary, mitigating the risk of a potential client. That's what we built.
That's super interesting. It sounds like you're recognizing patterns within an industry and that you don't need to go, necessarily, to the person who's receiving that case, but also to others in that industry to head it off the pass and get ahead of the game.
Right. It could be retail, manufacturing, hospitality, life sciences. I just used auto as a simple example.
That is absolutely fascinating. The approach that you took can be also applicable in many other areas, not just the legal profession. You could be looking at inbound marketing messages and trying to better uncover and qualify leads from that perspective. I imagine the approaches you took were predominantly in the NLP area as well.
I'm curious. Was there an actual increase in the number of clients landed or anything like that?
Yeah. I do know those numbers. I can tell you that the accuracy of the model was over 96% when we rolled it out. I love telling this story, too.
Once we went live with it, my friend Amy wrote me an email, and the email said something like this. She said, "All right, Peter. I have got to be honest here. When you said you could do this, we were more than a little bit skeptical, but this thing has knocked our socks off."
A couple of months later, it would've been November of 2019, just before COVID, that my friend, Amy, presented a high level discussion of the solution at a legal competitive intelligence conference. Thankfully, she shared some of the comments that came back, and it was really great stuff. Peers were saying, "My God, you guys are at least five years ahead of every law firm."
Another one was something like, “This seems like a license to print money.” So, right on that aspect, we're looking at about $20 million per year. We're a billion dollar law firm. So, call it a 2% increase that this does, year over year. The gift that keeps giving. It's been helpful.
This is fascinating. I do have to say, Peter, that those in the audience who are not in the legal profession are probably nodding their heads and thinking, "Eh, we've done something similar years ago, in my own company." I just look at their day to day and hear stories about some who are a bit behind the rest of the industry. I think it's due for some modernization, certainly in the data science role.
I wouldn't argue with you at all. That was one of my, I guess first ‘aha’ moments too, of coming to a law firm, was, "My God, they're at least five years, maybe 10 years behind what I'm doing for other industries." I'll just call them Fortune 1000. Part of it's really nice. A lot of those use cases that I was able to do earlier in my career over those 20 years, I'm bringing them to a new vertical, and all of a sudden I'm being celebrated.
More importantly, too, I love the opportunity to speak and share what I've done. There's that fine line. You have to protect the secret sauce, but at the same time, sure, I'm not giving any modeling techniques away. So, we're just talking about generalities or possibilities here.
If you're a partner in a law firm listening to this podcast, you would think that this should be an area of investment. My father would lament having to go in front of certain judges just because he knew what their tendencies were. The way he found out those tendencies were typically just word of mouth. It was just, "Oh, I went and I immediately talked to some of my fellow lawyers who had pleaded a case in front of this judge and tried to understand what I was up against."
If you're building these baseball cards, if you're building around these various judges that you might be seeing, I mean, that totally changes the game. That takes the bias and incomplete information that you're getting from your fellow lawyer and puts a whole new lens on it.
I think the other thing, Dave, to consider is —and you're absolutely right with that analogy from your dad—that COVID changed everything. Our culture at the law firm was that you need to be in the office every day for that exact reason. Very collegial, being able to talk to your peers and get their insights. COVID destroyed that. So, what do you consider, or how do you keep that competitive advantage when you're working out of your kitchen for the last year and a half? So, data and analytics maybe fills that gap.
I'd love to know what you think is next in the world of legal analytics. What are you thinking about a few years into the future? Where do you think this is headed?
One is we're starting to see it in the industry. Traditional big four consulting firms Deloitte and EY are getting into the legal game. So, what I sense is, not too far in the future, clients are asking to bring in accountants and traditional CIO advisory and legal to work as one consolidated team together on their business problems. I think that's a real trend that we're going to see continue.
Secondly, in the lawyers' professional responsibility, the American Bar Association basically talks about that. Attorneys have a, the word is duty, to keep abreast of technology. We're still kind of early in that wave from a legal perspective, but I would foresee in the next two or three years, we're going to see our first malpractice case because some law firm represented a client, they didn't use the data and analytics, the ruling went against that client, and they're going to sue them for professional malpractice. So, those are my thoughts, but I see the writing on the wall. It's going to happen and they have to be able to bring it to bare.
I wonder if there's an easy way to appeal cases just by using some of the analytics that you provide. So, if you have paralegals who did something manually, and then you are able to take a look and say, "Actually, what they calculated as—I don't know, the number of contracts that are overdue payment or whatever—is not actually correct, it's different.” That materially changes the case in some way. I don't know. That would be an easy way to audit these things.
You could look at all the cases that went to the appellate courts and then you could identify which ones were successfully overturned. If there's any patterns there, same thing. What were the arguments made in this district? What seems to resonate with them? I've never taken it, thinking through the appellate process, but I think the concepts are interesting.
I've learned a lot today. It's very interesting that the world of data science can be applied in all sorts of different realms. The legal world is, I think, due for a little bit of a refresh when it comes to data science. I wish our judges were out there, not biased and just read the law as is. There's a lot of interpretation and there is a lot of that gray area. Understanding what the judges' tendencies are, so that you have a better chance of winning the case, makes a lot of sense.
I think helping out with business development makes a ton of sense, even helping out with recruiting and HR. I mean, these are things that every company has to work with, whether you're in the legal profession or not. One next step for those in the audience out there might be just to have a conversation, like Peter did internally, with your head of council. Talk a little bit and try to understand their problems and some of the manual processes that they're working on. See where data science can play a role.
That last point, Dave, is absolutely critical. Stepping outside your comfort area, which is maybe just data and analytics, but having those business discussions. Find out about your internal clients. What's keeping them up at night? What are their processes? Can any of those processes be automated or improved? I mean, these are great ways to endear yourself to the business.
So, Peter, the world of legal analytics is certainly new. It’s what spurred me on to have you on the podcast. Are there others out there that are starting to recognize this as its own thing? Are you starting to see an industry form here around legal analytics?
A lot of startups in legal tech and a lot of venture capitalists are pouring money behind them. So, I think there's absolutely a market there. From a personal standpoint, I kind of tease my wife. It's like, "You better be nice to me. In certain circles, I'm a big deal." And she laughs, but it's really true.
I would say maybe from 2019, I've been asked to speak at conferences, whether it's to be on a panel or perhaps even keynote, been written up in a number of legal magazines, like Law 360, American Lawyer. Most recently, I was named in the top 100 innovators in data and analytics. I was a finalist this year for a legal innovator of the year.
So, yes, I guess the recognition is coming. It's really a credit to the firm and the leadership to give me the ability to have a canvas to paint on, so to speak. People are taking notes that we're doing some really innovative things at the firm. This is not only from a business side, but as we've talked about, to help our clients.
Well, unfortunately, thanks to this podcast, hopefully the secret will be out and you might have some competition out there. But as they say, if you didn't have competition, that might be a bad thing. But with competition, there's recognition that this is an area where investment needs to be made and, hopefully, you'll be seen as one of the vanguards, if you will, one of the innovators that spurred on the world of legal analytics.
Peter, thank you so much for being on the Data Science Leader podcast. I really appreciate you taking the time. If people want to get in touch with you, can they hit you up on LinkedIn?
Awesome. Well, thanks a ton to Peter and have a great rest of your week.
Listen how you wantUse another app? Just search for Data Science Leaders to subscribe.
About the show
Data Science Leaders is a podcast for data science teams that are pushing the limits of what machine learning models can do at the world’s most impactful companies.
In each episode, host Dave Cole interviews a leader in data science. We’ll discuss how to build and enable data science teams, create scalable processes, collaborate cross-functionality, communicate with business stakeholders, and more.
Our conversations will be full of real stories, breakthrough strategies, and critical insights—all data points to build your own model for enterprise data science success.