In a prior post on IBM Watson, I noted that Legal OnRamp uses Watson “to process and understand high volumes of contracts.” I describe here more about this and provide context about Watson and contract analytics.
Legal OnRamp Uses IBM Watson to Analyze Contracts
OnRamp describes itself as offering “hosted software and services to help legal departments deliver results more efficiently with higher quality and lower cost.” I recently spoke with Paul Lippe, CEO (and former GC of Synopsys, an AI software company for computer-aided design), and other management to learn more about their use of IBM Watson.
Borrowing from IBM Watson’s notion of “cognitive computing,” Paul calls contracts the “cognitive backbone” of the enterprise. Yet he points out that companies rarely characterize or track contracts well. On a day-to-day basis, information from contracts drives most business operations: revenue recognition, compensation, services and product delivery, risk assessments, lots of R&D and IP asset creation. But while companies put enormous energy into contract negotiation, they typically put the final documents “in a drawer.”
When major corporate events or deals happen – for example, M&A, divestitures, securitization, resolution planning (Dodd-Frank), and new revenue recognition rules – companies hire professionals to do some level of due diligence on their contracts, but with a narrow risk and exception focus. With better tools and methods, companies can turn contract details into data and ultimately “code,” better managing both risk and revenue.
That is where Watson comes in. The common view of Watson is as Jeopardy champion or Q&A machine. Paul reports that some firms are “exploring this ‘encyclopedic’ approach of winning [legal] Jeopardy”. In contrast, Paul and the OnRamp engineers – who are getting help from Prof. Dan Katz in this project – are realistic about how Watson best helps do legal work today. They see machine learning as “an additional member of the team,” supporting and scaling the work of lawyers in large projects.
According to Paul,
“many law firms are eager to try something in the realm of AI or Watson. We’ve created ‘Watson-powered deal rooms’ with some of our most sophisticated Silicon Valley customers, which are now being adopted by firms. Firms can improve their execution of existing deals and start to bring together different elements – project planning, tighter collaboration with clients, maybe fixed fees – in large matters, while getting in on the ground floor with Watson.”
By making it easy for firms to implement Watson, Paul thinks OnRamp will have between 12 and 20 firms using Watson by year-end. He says, “we’re batting 1.000 when it comes to selling Watson to managing partners, because they see it as a catalyst for a number of changes they want to accelerate.”
So, just how does OnRamp use Watson? It feeds a collection of contracts to IBM Watson and other machine learning tools to automatically answer some questions and accelerate the human review process on others. Because OnRamp works directly with companies, and has already done large scale human reviews, it has access to far more contracts than any law firm. That means OnRamp can build on its efforts over time, enhancing its understanding of contracts. The article IBM Watson: It’s (Almost) Elementary (Metropolitan Corporate Counsel, 12 May 2015) says it well:
“When the OnRamp team completes the heavy lifting of deconstructing 10,000 contracts, Watson jumps in and makes digital mincemeat of the other 90,000.”
OnRamp’s goal today is to deliver a 20% cost reduction for a law firm in a typical diligence exercise and “enable companies to deal with scale in ways that were previously impossible, simplifying resolution planning or revenue recognition transition”. OnRamp’s machine learning doesn’t aspire to substitute for human judgment, but is much faster (and more accurate) than a lawyer in, for example, seeing if a particular 79-word cluster that has already been tagged as “Indemnity Clause Variant 6” is repeated. So it will replace a significant amount of lawyer labor. Paul reports that Watson should substitute for more labor with each subsequent use, but agrees we should track that closely to see if and how it happens.
Legal Market Applications of IBM Watson
My recent post Legal IBM Watson: Business Model and Reasoning Modes reported two other IBM Watson legal applications: ROSS provides a legal research and UK law firm DWF has built an internal work allocation system.
In that post, I suggested that building a Watson legal Q&A application is likely a big and expensive project and that the business model to do so is unclear. Contract term extraction may not be as glamorous as Q&A. For the moment, however, it seems the one with clearest business case, and a way for firms to show clients they are serious about innovating, kicking off win-win conversations with clients.
Other Approaches to Analyzing Contracts
OnRamp is not the only company offering contract analysis service. Legal process outsourcing (LPO) providers (and I include Axiom Law here) have, for several years, offered similar output.
Some providers use Seal Software to conduct a first pass term extractions. Separately, Kira and eBrevia offer automated approaches for conducting due diligence. My March 2014 post, Improving Contract Management and Analysis, describes Seal and Kira (then called DiligenceEngine) in more detail. KM Standards works with smaller contract sets to find common provisions.
I suspect there are other contract analytic tools I have missed. But as Paul points out, Watson’s R&D investment is probably 100x all these companies combined, and so has the potential to ride a much steeper performance curve.
Paul and I have discussed legal tech, innovation, and “disruption” for years. Every time I complain how slow it is, he points out that providers do not need the entire market to shift. According to Paul, “Watson is potentially the most ‘disruptive’ technology we’ve seen in law, because it is being adopted across clients and because it will create relative competitive advantage for those law firms who figure out how to use it.”
These days, I shy away from “disruption”. Whatever the characterization, I expect to see continued growth in automating high volume processes. As many commentators point out, automation, especially artificial intelligence like Watson, is more about augmentation than outright replacement. At least for now.