EPISODE 208 – HOW MACHINE LEARNING IS BEING USED FOR CONTRACT ANALYSIS

We may not yet be at the point where robots are taking all of our jobs … But why not pass off some of the workload at least to machine learning artificial intelligence software that sets the stage for users to employ its contract analysis and helps attorneys in their due diligence. The Director of Marketing of Kira Systems explains how it works and who it works for. Listen to the podcast below or download it for later. Make sure to subscribe, share and review the podcast.

 

Podcast: Play in new window | Download

Tracy: This is Tracy Merda and today I have a special guest Robin McNamara, Product Coordinator at Kira Systems, a leader in helping enterprises uncover relevant information from unstructured context and related documents. Hello Robin. Thank you so much for joining me.

Robin McNamara: Hi Tracy, Thanks. Thanks very much for having me.

Tracy: As someone who spends a lot of time myself looking over paperwork and contracts and things like that. This topic that came up today with Kira Systems is absolutely fascinating. I’ll be honest with you. It sounds too good to be true, to be quite frank. So tell me what is Kira Systems and how does it work?

Robin McNamara: We are a software company that assists law firms with identifying and analyzing important clauses in contracts as well as other important documentation that they might have. In particular we assist or the genesis for our company was assisting law firms with the due diligence process associated with Legal work which is a process in which usually large corporate firms use huge volumes of contracts in order to identify risks or other information that might be relevant to a transaction thereby affecting the purchase price or how the transaction is structured or whether the transaction can really go ahead and happen at all. So this was a process that our founding team, our CEO Noah Waisberg and CTO Alex Hudek identified as an area of legal work that could be sort of improved in the way that it was conducted in the past. They saw an opportunity there and we built a software team to assist with that.

Tracy: So how long have you guys been around?

Robin McNamara: Noah and Alex founded Kira Systems in 2011. Noah’s background is as a corporate lawyer at the firm called Weil, Gotshal & Manges in their New York office where he did a lot of this legal work called due diligence which basically involves reviewing usually large numbers of contracts as I said before, in order to identify the risks and also test opportunities and so he did a lot of this work and I think the characteristics of that sort of work really lent itself well to the type of technology that we use which is where Alex comes into this. He has a PhD in Computer Science but specifically has a specialization in machine learning from a school called Waterloo which is sort of Canada’s MIT. So the two of them got together and were able to come up with a machine learning based solution to this very common legal problem associated with due diligence. Due diligence being a task that is pretty difficult to do well. So really smart and talented people working at prestigious law firms often struggle to do it well because the information you are trying to find is potentially hidden or nuanced in such a way that it is difficult to recognize. But it’s also fairly a repetitive process so oftentimes you are looking for the same sort of things again and again. So in their combined expertise they were able to come up with a software solution that delivers both time saving as well as accuracy benefits and for that reason, our company has been able to attract some very high profile clients including some of the largest and most traditional law firms in the world like DLA Piper, Clifford Chance and Freshfield as well as some of the other large professional service organizations like De Voight. That is a brief history of our company which came to be in the last 5 years or so.

Tracy: You guys came at it with both the technology side of it as well as having someone with firsthand experience and probably a lot of appreciation for anything that can ease the process of going through contract after contract, I imagine.

Robin McNamara: Yes. Exactly. It’s our dual expertise that is really important. We seek to provide a good solution in this area. We have a number of computer science PhDs in our staff at the moment as well as a number of really talented developers and designers. We also have quite deep legal expertise particularly in our primary target market. So we have a number of people on our staff who have a lot of experience doing MNA work in a number of different contexts and we think that it really gives us both a solid understanding of what our customer’s needs are as well as the cutting edge technical expertise needed to do really exciting and interesting things in this space.

Tracy: Let’s talk a little bit more about the technology behind Kira systems. Tell me a little bit about how it works? Obviously I am by no means an expert in this field so if you could dumb it down for me in layman’s terms. What is it looking for when it’s going through documents and things like that?

Robin McNamara: There are a few different ways that you could approach a problem like this. A problem like this being where you have a lot of unstructured information and you want to find certain phrases or words that are suggestive of broader concepts than you are looking for. Systems in the past used oftentimes rule based approaches to solving these problems. You might come up with a recipe that looks for certain keywords, maybe in a certain order or something like that that you can sort of as a human who is very well experienced in a given area might be able to construct a recipe to find a lot of examples of a certain type of contractual provision. An example of that would be something like an assignment provision usually uses the word assignment and maybe a handful of other words and therefore if you search for those words in close proximity to one another, then maybe you are going to be able to capture a majority of the instances of an assignment provision. The significant difference here between that sort of approach which had been tried a bunch of times in the past and it was moderately successful as a solution to this sort of a problem but had some really significant drawbacks. The difference between that and what Kira Systems does is that we use technology that really determines that recipe for you. So we have technology that will basically break down natural language into it’s component parts which we call features and then on it’s own will come up with an appropriate recipe to weigh the different features to find the information that you are interested in finding. So for example, an assignment provision, although you as an expert in the field might know that it usually contains the word assignment. There are all sorts of other features that might be relevant to making that determination that you wouldn’t be able to maybe pick up on or articulate well. So maybe the assignment provision occurs generally in a certain position within the contract or within only certain types of contracts. Potentially the sequence of words used in assignment provision is atypical. So there are all sorts of things that it would be very difficult for a human to both pick up on and articulate well and oftentimes unpredictable things and for that reason, if you can have a system make those pro se determinations in a mathematical way where you end up having happened is that the system is more accurate and importantly a lot more robust to unforeseen types of phrasing. So if someone uses synonyms or uses a new sort of phrasing, our machine learning system is able to accommodate that and maybe identify that it is an assignment provision even if it is not phrased in exactly the same way as the system has seen before. Does that depiction make sense?

Tracy: Absolutely. This seems like a no-brainer to me. Are a lot of law firms using these systems? What else is out there that they are using instead of this system? It seems crazy to me if someone doesn’t jump onboard with this today.

Robin McNamara: As I mentioned, we have a number of very significant law firms and corporations and other professional services firms using our software but this base in general, is fairly new. There aren’t too many companies that are approaching contract review using machine learning. There are others but we are all fairly recently established companies. If you look at the way this space is changing there is fairly rapid adoption of Kira Systems technology in law firms as well as corporations and other professional service firms. I think the main reason for that is these organizations are coming to the conclusion which you just mentioned, which is that if you can do this more accurately and in less time then it’s a no-brainer that they should use this to enhance the way that you are doing this work.

Tracy: So you mentioned that it’s in it’s infancy slightly in terms of the companies being around and things like that but you obviously, I’m sure are moving at a fast pace. Where do you expect to see Kira Systems in the next few years as to maybe 10 or 20 years down the line?

Robin McNamara: It’s not to say that we’ve definitely grown quite quickly. We are about 40 employees now which is a good number for a fairly recently established company. We are at the moment drawing a lot of attention for the core use case or the reason for our company being established around due diligence and improving the way that it is done at the moment. I think that there are a number of exciting things that we see on the horizon and there is still around this high level goal of improving the way that contract review is done. But there are a number of different ways that we can assist with that process. An example is that we are excited to assist with more is the sort of knowledge management around best practices for drafting contracts in different sorts of ways. At the moment we focus on primarily with assisting the contract review but we could and have more recently been doing some quite sophisticated studies around deal points and what the trends are in industries with given types of contractual provisions. These are the studies that happened down in the past highly manually but the difference with using technology like Kira for a task like that is that you can do it much more quickly and therefore you can do a much larger volume. In early 2017, for example, we would be able to give legal professionals a fairly comprehensive understanding of the way that certain types of contracts have been drafted in a given industry in 2016 whereas in the past, getting that sort of information would take a long time and be very expensive to have lawyers come in and construct those reports manually. So we see a lot of opportunity to augment the industry knowledge around best practices for contract drafting as well but there are a number of other ways that contract review or creation or a bunch of things around contracts could be improved and our system works either with it to work with. In fact a number of companies that you’ve also had on this podcast in interesting and exciting ways. So also short of ensuring that our software plays well with the sort of adjacent products in this space is another thing that we are excited about.

Tracy: I love this and I’m so looking forward to seeing what you people do with this especially for a lot of the attorneys that we work with; I already have 5 in my head that I know this would help with. So if someone would like to contact you Robin for more information and see if maybe your services are a fit for their firm, what is the best way to reach out?

Robin McNamara: We have a website as most companies do which is kirasystems.com. There you can read about our company and our product as well as a bunch of literature that we’ve produced about this sort of industry and our ideas on how and why machine learning is a good way to approach these sorts of problems and you will also find our contact information there. So get in touch with us and schedule a demonstration if you are interested in learning more.

Tracy: Alright. Well, I just want to give a special thanks to Robin McNamara, Product Coordinator at Kira Systems. Robin Thank you so much for this great information and for your time today. I know you will have a lot of interested listeners out there

Robin McNamara: Yes. Thanks very much for your time as well Tracy.

 

Podcast: Play in new window | Download