Legal tech made simple

Interview with Tim Pullan, founder and CEO ThoughtRiver

April 16, 2020 Dom Burch
Legal tech made simple
Interview with Tim Pullan, founder and CEO ThoughtRiver
Chapters
Legal tech made simple
Interview with Tim Pullan, founder and CEO ThoughtRiver
Apr 16, 2020
Dom Burch

Tim Pullan is founder and CEO of ThoughtRiver, the intelligent contract pre-screening leader.

In 2015 Tim relocated from Singapore to UK to develop ThoughtRiver’s technology in Cambridge, one of the world’s leading centres for natural language processing and machine learning.

Tim is a regular writer and speaker on the role of data and machine learning in transforming legal risk management. 

In this podcast he tells us about ThoughtRiver and how it is helping lawyers use their time more efficiently rather than constantly reinventing the wheel. 

In Tim's view once you have a knowledge-tree like structure in place and introduce consistency into the common parts of contract review, you can then start automating things, be that remediation, or reporting and information sharing. 

But the thing that enables all of that to happen is to have a consistent universal data model on which all of these other use cases and applications can rest on.

AI then becomes an enabler for the delivery of expert advice to the user to help them speed through the task of creating and approving a contract. 

Show Notes Transcript

Tim Pullan is founder and CEO of ThoughtRiver, the intelligent contract pre-screening leader.

In 2015 Tim relocated from Singapore to UK to develop ThoughtRiver’s technology in Cambridge, one of the world’s leading centres for natural language processing and machine learning.

Tim is a regular writer and speaker on the role of data and machine learning in transforming legal risk management. 

In this podcast he tells us about ThoughtRiver and how it is helping lawyers use their time more efficiently rather than constantly reinventing the wheel. 

In Tim's view once you have a knowledge-tree like structure in place and introduce consistency into the common parts of contract review, you can then start automating things, be that remediation, or reporting and information sharing. 

But the thing that enables all of that to happen is to have a consistent universal data model on which all of these other use cases and applications can rest on.

AI then becomes an enabler for the delivery of expert advice to the user to help them speed through the task of creating and approving a contract. 

Dom Burch:   0:09
Welcome back to��Legal Tech��Made��Simple with me, Dom��Burch. I'm not a lawyer and I'm not a techy, which makes me perfectly qualified to make legal tech simple. Join me in this podcast��as��I interview legal engineers, software developers, law firms, large corporations and companies��who are��at the leading and cutting edge of legal tech. And I'm delighted today to be joined by Tim Pullan. Tim is the founder and CEO of ThoughtRiver. They describe themselves as the intelligent, contract pre screening leader. In 2015 Tim relocated from Singapore here to the UK to develop ThoughtRiver's��technology in Cambridge, one of the world's leading centres for natural language processing��and��machine learning. Now Tim's a regular writer and speaker on the role of data, and machine learning in transforming legal risk management, and we welcome��Tim��to legal tech made simple. Tim delighted to have you on our podcast,

Tim Pullan:   1:02
Dom it's an absolute��pleasure. Thank you very much,��what a great intro!

Dom Burch:   1:06
I mean I'm in esteemed company. It's one of the delights of having this podcast. I'm getting some big hitters on the show already, so let's jump back a few years shall we? And just tell our listeners what prompted you to create��Thought��River?

Tim Pullan:   1:19
Yeah, I think like a lot of��entrepreneurs that are��driven by a kind of vision, it was just kind of an itch��that I��just had��to scratch. Um, lawyers were basically��really get negotiators. But they were actually very poor recording what they had negotiated. And I��decided, just through many years of working in law and outside of law, that the answer to that was ultimately to bring structure to, at least the majority of what��lawyers do.

Dom Burch:   1:55
I'm right in thinking that your��own research has identified huge inefficiencies,��particularly around contract review. I think you estimate that to be around $40 billion. So just explain for people who perhaps on au fait, yet with��ThoughtRiver how your tech helps reduce that wastage.

Tim Pullan:   2:11
Yeah. I mean, it's probably best to start with a problem,��because��$40 billion is a lot of money, but actually, it's��really a drop in the ocean of the real problem because that's just the cost of lawyers reading stuff they don't need to read. The real cost is businesses they're supporting, the deals they're doing, which are slower, more costly, opportunities lost. And��we haven't actually tried to calculate that, but it is orders of magnitude bigger. You can talk about this big macro numbers and people say��wow,��but��actually, what's really super interesting whenever you've got a macro��problem like��that is to��dive straight down into the micro.��You know what is the reason why we have this situation today, where lawyers are having to read things that don't need to read,��and in contract the micro is to do with what happens on every deal. Actually hundreds,��if not thousands of times, sometimes on every deal, which is that lawyers have a tendency��to reword things which are quite common in their own way. They do that with professional pride, but every time they do that, they are creating a job for someone else, which is to read that thing. Or actually, often many other people. And��that��in micro is the problem that gives you the macro. I'll put some real numbers on it in terms of that micro problem. So we got��our data��scientists to look��at just over 1.5m��contracts. And��I asked��them��to tell me how many different ways lawyers have found to��write��the clause, the��point of meaning in a contract, which is so the law of this contract is X, you know,��could be England��and Wales��could be New York could be��California. One of the simplest things you can��say in��a contract, average clause length is��50��or 60 words. And in the 1.5 million contracts, they found 335,000 different variations that��lawyers had created of that very simple thing in a contract. And then it's just one small example. But��if you just��take��that one��small��micro example and you look at, say, the population of lawyers in the states about 1.3 million,��pretty��stable number, and you made some very conservative assumptions around you know, they may only��encounter��1% of these variations in their entire careers for example, you end up with two million hours being wasted every year by American lawyers simply reading law of contract clauses. And of course, actually, there are many, many more complex��but��very common things in contract, which are far more vulnerable to being reworded��and��wasted reading time, so that, in micro, is the problem that we face today when we are trying to bring acceleration��and more structure and efficiency to the contracting��process.

Dom Burch:   5:06
It is sometimes difficult to visualise a product like this. But just describe, because I guess what you're getting down to is the fundamentals here, is to build playbooks and common clauses. And then��to��have this library almost like a tree, if you like a decision tree that allows people to build a contract with standardised terms and obviously tailored for the specifics of that organisation or that industry. But��is that how in the background, is that how the product's working?

Tim Pullan:   5:33
Yeah, I think there's just one��mini��step between the problem��and��the playbooks, which is structure. So all those thousands of ways of saying the same thing,��fundamentally the first step is we summarise it��as��one thing. And we do that in order to enable lawyers to assimilate information more quickly. So instead of having to��read��lots of different variations, they only ever get a kind of consistent summary of the same common issues. And that's��a fundamental foundation to our approach to the problem.����Once you have that foundation, that��knowledge-tree like structure that you referred to,��that introduces consistency into the common parts of contract review. Once you've got that, you can then start automating��things. You can start automating��the application of playbooks. You can start automating remediation. You can start��automating reporting and information sharing, but the thing that enables all of that��to happen is to have��consistent universal data model on which all of these other use cases and applications can rest on.

Dom Burch:   6:43
So bring it right up to today,��describe for us maybe��a use case of a large organisation, a lot of companies that we're doing work for, will be working on third party paper. They're not necessarily contracting on their own terms, but they're having to, you know, do deals with other people's terms and having to read those contracts line by line to see there's nothing in there that they don't like. Or there's, you know, important clauses that��are missing.

Tim Pullan:   7:09
Yes, a classic use��case��is in the commercial contract space.��So if you take one side commercial contracts, which would be sales contracts.��As you say, a lot of our clients, even though they're big companies they trade with other companies that may be��even bigger but which insists on their terms.��So a typical portfolio might be, you know, 40% own terms but amended, which is just as bad as being off third party��paper, or the starting��point was third��party paper. Take a sort of a legal team that's supporting��a sales��function so they get inbound contracts. It might be the first draft of a contact. It might be the seventh draft. You know that they have��been��working��on with their counter party. Before they've even��seen this draft, what's happened to it is that it's��already��been��reviewed by our system,��and��so then it's sent to them by email.��They get an email in the same way they would normally on Outlook. But when they open that email the��documents attached to that email, first thing they see in the email is��a summary of what's in that contract. So they've got it a sense of the size of the number of issues that they may need to deal with that kind of nature of the task. But when they open the word document, what they have is in the right hand panel. Embedded within the Microsoftware environment is they have a commentary, which shows them where the��sort of the normal, the routine things that they need to look for��are��set out in that contact if they are issues. And��where there are issues, it's only showing them the text which relates��or carries that issue.��Or maybe the issue is that there's no text in there, as in there's��a clause��missing. But then it also helps them actually do their work. So the work is, well��I either accept this so it enables them to��take��action like accepting, or I need to redraft it. So��it��will red line the document using native word functionality or comments on the doc. Maybe they need to put a comment in there for an internal stakeholder or indeed, the��counter��party. So it's helping them to��do all��three of those tasks and then helping them get it back out the door as quickly as possible. Now, actually, even though��these sales��agreements are pretty, sort of, always��selling the similar thing. Actually,��in every deal there'll be something that's different, and that's really where you want your lawyer. You're experienced negotiator focusing. So what we are doing is we're helping them deal with the more common things and then having them spend more of their time��dealing with the the issues which are actually more important.��Which generally��don't sit in the legal terms��& conditions,��generally those issues sit in some of the sort of commercial arrangements that really define whether this is��gonna be a��good or bad deal. Of course,��their��manager can see what's going on across on all of the user base because each one of those contracts that's coming in each version of that contract is generating huge amounts of data that��can then be��used for insight��and analysis to help kind of fine tune��processes.

Dom Burch:   10:30
And as I understand it, you have a full service API don't you? So your product is one that can plug in because increasingly, what we find when we're talking to customers now, is that actually we're either having to do a bit of a��Frankenstein job and take some new tech and help it plug into all the other things they have. Or, frankly, don't give them any new tech. Get this process sorted out first and use��the tech they already have and actually��Office��365 is a pretty powerful tool isn't��it? When you use it to its full degree.

Tim Pullan:   10:59
It is for me. It is King.��I started out in law mid nineties and mid nineties we used to red line documents in Microsoft Word and we used to��send them��to our��counterparts in��Microsoft��Outlook, and��25 years later, most lawyers, 95% of lawyers��who are doing contract negotiations, do exactly the same thing. And there's lots of reasons for it. It's not just to do with lawyers sort of comfort zone. It's actually to do with the design of these tools, which is so fundamental, it's so well designed for��the��sort of work the lawyers do. So��for us ��it was essential that everything that we do, bringing all the intelligence and support we needed to bring it into the environment where the��lawyer��felt, you know, not just most comfortable but most concentrated, most effective,��most productive. And so we see our role, yes, the API you��can plug us into��any host system, and indeed, you could plug us into other word processing systems as well. But fundamentally, we believe that lawyers, certainly in what they are producing today, being��contract review, are��most effective when they're working in Microsoft Word.

Dom Burch:   12:14
Now, I'm right in thinking you segment the contract review universe into two��buckets. You might not call them buckets, but anyway,��one��would be intelligent searching the other intelligent, summarisation with companies like Kira,��SEAL, eigen on one side and you on the other. So just give us a sense of what makes you different.

Tim Pullan:   12:31
So if we if we take the��Kiras, the Luminances and so on and so forth. What those guys really, where they come from is a world where they are reviewing kind of closed bundles of documents, typically in due diligence cases, so unsurprisingly��their early markets started with��law firms and legal service��providers, and that's really the origin of their products. And essentially, what they are enabling is an intelligent search capability, which is��obviously��very effective for those��kind of use cases. Where we��draw quite a clear distinction is between that��and��the market in which we sit, which is intelligent summarisation��which was really driven again by the use case. So intelligent summarisation��is like taking AI and��attaching an expert system to it.����Because if you are gonna do the first��pass review of��a��contract, what��the user really needs to����understand is not that there are some��clauses that are gonna look like this clause��in there. But��they have a sense of what the fundamental points of meaning are in that document��and��the relevance to them. So there's always a layer of expert sort of��interpretation,��that needs to��sit on top of AI. Actually, AI just becomes an��enabler for the delivery of expert advice to the user to help them��speed��through the task. So��very different technologies and��very different��stacks. If you were to summarise it briefly, you'd say that intelligent search really is originated in retrospective review use cases where you're looking back on historic contracts, whereas the kind of intelligent��summarisation capability��has really being designed for a world where you're��pre signature. You're in the process of negotiation, and that really has driven the design of��solutions like ThoughtRiver, which are focused on that use case.

Dom Burch:   14:27
In the current climate, obviously and the crisis we're going, through.��It's gonna make it almost impossible, I guess��for��people in the sector to carry on doing what they've always done. What do you hope the positive outcomes will be in the months ahead?

Tim Pullan:   14:38
First of all I hope everyone remains safe and well and that, you know, people find a way to adjust their businesses and the way they're working. As we��have had��to��adjust to the��new realities of the world, but I think that stepping away from it a bit. If you look at what's happened in previous crises like the great financial crisis, actually, these are times of opportunity to bring about change, which is often��been��coming for a number of years. And I think that certainly in legal services and corporate legal services, I think that what we'll see is an acceleration of, ��not just technologies but new kind of ways of working which ��deliver the kind of benefits that particularly Corporates are looking for in processes like contracting. I see this is an opportunity��for��innovation and change. I think, you know, there will obviously��be some hard times, but I think that it's actually��if anything you look at previous crises, it has been��a catalyst for positive change overall. So I very much expect to see that this time around, possibly up to 12 months, depending on on on how deep, you know,��the disruption is��not just from the��pandemic, but from the impacts of��the pandemic. You know,��in terms of economic destruction.

Dom Burch:   16:12
Well,��Tim is��has been an absolute pleasure. Thank you for taking the time to join us.

Tim Pullan:   16:15
Thanks Dom.��Been a pleasure. Thank you.