17:10, September 01 547 0 law.com

2017-09-01 17:10:06
Former Wal-Mart Head of E-Discovery Dishes on His Move to AI-Startup Text IQ

Aaron Crews, general counsel and vice president of strategy at TextIQ, photographed in their Manhattan office on August 30, 2017.

Photo: David Handschuh/ALM

Aaron Crews discusses how his experience at Wal-Mart motivated him to become general counsel and VP of strategy at Text IQ, and why legal needs AI now more than ever.

As one of the biggest private companies and employers in the world, Wal-Mart Stores Inc. has to contend with an enormous amount of legal actions every year. For many, just keeping up with the variety of cases, which run the gamut from tort actions to class action law suits, can be formidable.

But for Aaron Crews, who for three years served as senior associate general counsel and global head of e-discovery at Wal-Mart, the core challenge often comes down to one hurdle: effectively managing and understanding ever-growing relevant data sources.

Crews has steered Wal-Mart’s legal department through taming and reviewing more information than most companies handle over the course of an entire year. In August, he became general counsel and vice president of strategy at Text IQ Inc., a legal technology company offering artificial intelligence (AI) powered e-discovery, legal analytics and compliance platforms. He sat down with Legaltech News to discuss how his experiences at Wal-Mart motivated him to join the startup, and why he thinks legal needs AI to survive.

Plugged In

LTN: How did you end up at Wal-Mart?

Crews: At Littler Mendelson, I was national coordinating counsel for e-discovery for Wal-Mart as one of their outside lawyers, and eventually they convinced me to come in and run e-discovery for them internally. So in 2014, I made the move and became senior associate GC head of e-discovery.

LTN: What e-discovery challenges did you face at Wal-Mart?

Crews: Because of the scope and the scale of Wal-Mart, there are very large class actions and some long-running investigations that involve just ridiculous numbers of people. So just the custodian identification, data collection, figuring out how to get through all that data is a tremendously difficult process.

Wal-Mart had multiple matters where the number of people on hold for that matter would dwarf what a lot of companies put on hold for an entire year, for all their litigation combined.

LTN: How did you deal with all these legal holds?

Crews: It really was a people, process and technology kind of thing. We took some existing technologies and made them better, either working directly with the vendor or working with our IT group. It’s not abnormal for my team or I to write out pseudo-code and then say, ‘OK here’s how it should work, how it should function,’ and then we would have people help us build it.

We also built an entire team whose whole job was researching and drafting legal holds, sending out reminders, auditing compliance for legal holds, and releasing them at the end of a case. There is an entire organization we built that was designed to do just that, and it was the only way we could handle that scale.

LTN: Given the scale, how did you rein in e-discovery costs?

Crews: A very big piece of my mandate while I was at Wal-Mart was to redesign how the company approached litigation, and so we built a brand new process from scratch. In the midst of that, I came up with a concept I called ‘core discovery.’ The idea of core discovery is that in any matter, there are a handful of individuals who are the center of things, and in any litigation or investigation, you know who those people are right out of the box, because they are mentioned in the compliant.

We would use tools offered by Brainspace and Text IQ and some others to start to identify documents out of those custodians, and we would do a social networking analysis to figure out with whom those people were communicating about things that mattered [in the case or investigation] during the time period we cared about.

Eventually, what you see is a social histogram of everybody who is talking about things that matter in your case. You have a complete map of it. This gave us a very defensible basis for cutting off custodians and things like that.

We would also streamline the review process by using document reviewers instead of outside counsel to do the first cut and have outside counsel look at the product of that process.

What you saw what was I call a “virtuous cycle,” where outside counsel were asking the questions and dictating what was being reviewed, and then they were consuming the output of that search and using that to further refine their questions. That process really shortened the time frame and dramatically cut the spend that was involved.

LTN: What made you jump ship to Text IQ?

Crews: [Implementing Wal-Mart's new review process] was one of the pieces that caused me to take my current Text IQ job, because I’m so impressed with the technology and what Text IQ was doing, and realized that, right there, that is the future.

I came to Wal-Mart to try and change how people litigate. But realized if I go to Text IQ and help this company do what they’re doing, I’ll have the ability to really help reinvent how review processes work by giving people the right technology and tools.

LTN: Why do you believe AI is so pivotal for legal?

Crews: The problem that is colloquially known as e-discovery is really a technology-driven problem, because what you’ve seen over the last decade or two since the late 1990s is essentially a trend whereby data volumes double essentially every two years. It is kind of a bastardized version of Moore’s law.

The only way to solve the problem, which is a technology-driven problem, is by using a technology-driven solution.

What AI machines do really well—which is brute force review of large data volumes, data pattern recognition and learning—and what humans do really well, which is rapidly synthesizing distinct amounts of information and making decisions based on that information—when you put those two things together, you get some really amazing outcomes.

If we don’t do this, if we don’t make this move to AI soon, my very real fear is that we will have started to price people out of justice. We will make it economically impossible for individuals or organizations below certain thresholds to get into an American court, because it’s just too expensive. I think it’s a real possibility for a growing numbers of people, unless legal starts to act like the rest of the economy—using normal business processes and integrating technology everywhere we can to improve the efficiency, lower the costs and improve the speed to outcome.