E-Discovery—The New Lean-In Moment

Which should suggest that people are overwhelmed, not in control, I suggest. “That’s exactly right,” says Dave, “if you use the same old technology. Trying to tackle discovery—especially e-discovery—without machine understanding is fruitless,” he claims. “You don’t have enough people, and you definitely don’t have enough time. It’s imperative to have technology that understands meaning, and understands what we’re asking for. It’s a need; it’s an imperative. It’s not even a ‘nice to-have’ anymore.”

But I’m still unclear on what pushed the tipping point over the edge. “Three years ago we worried about how to store everything. Open source indexing, Hadoop, etc. So people can now store everything. The question they’re asking now is: “What do I do with it all? How do I create transformative value from it?” asks Dave, somewhat rhetorically. “Now we can transform all this data into an asset. We call it a ‘brain that everyone can use.’ We’re still at the dawn of it, but we are getting there. We’re not quite at the age where machines are doing ?everything, but they are doing a lot of things for people.”

Getting the Hard Work Done

“You know that A.I. and machine learning are successful when they disappear,” says Dave. That’s a profoundly true statement. And Dave is a terrifically realistic person. “We now have the ability for machines to interact with humans. We are NOT,” as he underscored earlier, “at the point where all of the learning happens without people. But we’ll get there.” And it’s a damn sight greater than before. “We’re now able to allow people to do greater thing, and let the machines do all the grunt work.” Which is what all machines are meant to do, since Archimedes, I thought.

But the question hangs in the air: Who is responsible? Is it IT? Is it legal? Is it line-of-business? Where does the ultimate responsibility fall, and who is skilled enough in applying it and making it work?

“It’s becoming imperative for companies to find, curate and collaborate on information assets such as those, and that’s not just one person,” adds Dave. “Normally there are entire teams of knowledge management workers, KM officers, and in the case of discovery, teams of people who work together, both inside of the companies themselves and from legal service providers.”

“When people come to understand the importance, the conversation escalates dramatically. It’s frustrating now, because people work hard on information only to have it put away in some repository. You can—today—type in three words that you know are in the title of a document you created, and still not find it. That’s the frustration. I want to ask a simple question and get back exactly what I want. You should ?be able to take a paragraph from a document, and say ‘give me something else like this.’ We’re finally getting to that place,” claims Dave.

Moving Away From Transaction

Is e-discovery a noun or a verb? Is it something you constantly fuss with, and do again and again? (That’s the “verb” part.) Or is it something where you put content into a repository and walk away until you need it? (That’s the “noun” part).

“You’re hitting on something very important,” agrees Dave. “It’s constantly streaming. Even in e-discovery, the documents you’re looking for are changing every day. So it’s important to have a system that can learn dynamically, as new documents enter the system. We’re moving away from a transactional world. Now, there’s learning happening,” says Dave.

It’s very exciting and exotic, but there is no question that Dave has a grasp on what companies currently need to do.

That leads me to ask Dave one of my favorite questions, which is “Is governance something that can be taught as a policy? Or is it something that has to be applied as a technology?”

“It’s funny you asked me that. We think about that a lot. The industry is run by expert practitioners, and just like the data, that is changing all the time. It wasn’t till recently that there were even classes on discovery, and the ones we have are two years behind,” Dave laughs. But it’s important; the curation of content that leads to knowledge is a precious activity, and as Dave points out, it’s a moving target.

I describe my “arc” theory that most technologies go through. Gartner calls it “the hype cycle.” I believe that the first stage is “teach me about information governance.” The second stage is “show me how information governance can help my business.” The third (and final) stage is, “Let’s see how information governance can add value beyond the simple plumbing of repositories and storage. How can it propel my business?”

What I get from Dave is that we’re somewhere between stage two and stage three.

“There’s inherent value inside of information itself,” adds Dave. “We now have technologies that take us to a new starting place that’s well beyond the old manual way of doing things. In the old days you started with an index and a search term. Now, in this new world, before you ever push a search button, all the concepts in all your documents have been unified into a single intelligence. You have all the intelligence of all those documents at your fingertips. That’s a magnificent leap beyond where we’ve been before,” says Dave.

Dramatic, but true.

“The focus in discovery has changed. It used to be you’d sit down 50 lawyers around a (presumably large) conference table where they’d manually review documents. Then we went to search tools, where 20 lawyers could search with keywords from documents. Now, the focus is off ‘review’ and on ‘analytics.’ The sheer volume makes it impossible to review. What you want to do is get it down to as few relevant documents as possible. The job in discovery now is ‘How do I get down from 10 million documents to a few thousand documents that matter, as fast as possible?”

How, indeed. Legal discovery is a fact of life, whether it’s civil litigation or a merger/acquisition negotiation. I’ve come to accept that the ultimate answer to that question lies in text analytics. Call it machine learning, call it artificial intelligence, call it cognitive computing, whatever. It’s all saying the same thing. The bottom line is that technology tools are instrumental in providing the requirements for sculpting knowledge out of data. Get used to that.