Your single source for new lessons on legal technology, e-discovery, and the people innovating behind the scenes.

6 Questions with Text Analytics Expert Jeff Gilles

Andrea Beckman

Text analytics hit the e-discovery scene more than a decade ago, and in the past four years alone, usage of Relativity Analytics has grown by nearly 1,500 percent. The technology has been helping customers quickly gain insight into massive volumes of unstructured data and tackle everyday tasks as well as the industry’s most heinous projects.

Jeff Gilles joined the kCura team this month with more than 15 years of experience in software, eight of which have been in advanced text analytics and the underlying technology that supports a wide range of features in Relativity Analytics, and he’s ridden the tides of change in e-discovery along the way. We sat down with him to get his perspective on this evolution.

Andrea: What developments in text analytics most fascinate you, in e-discovery or other industries?

Jeff: It’s been an interesting and storied journey. Over the course of the past eight years, the technology has come a long, long way. The commitment to great customer experiences, particularly in the e-discovery industry, has led to fundamental changes. The addition of structured analytics like email threading, for example, was a major milestone in our history and the direct result of the market’s demands. Everything’s been tweaked and re-tweaked, and there have been countless enhancements and overhauls.

I’ve seen some very smart teams solve huge problems with text analytics, within and outside e-discovery. One example we’ve seen as a novel use of text analytics is tagging entities and asking the engine how those entities relate to each other—essentially letting the technology teach an expert something new about their own subject. This innovation has helped medical professionals discover new relationships between drugs and diseases, and it’s crazy what kind of impact that information can have. In contexts like these, analytics isn’t just helping a complex process run smoother—it’s game-changing.

What feature has been your favorite addition to Relativity Analytics and why?

This is tough, but if I have to pick, I’d go with the optimized training set option added to indexes, especially useful in Relativity Assisted Review projects. Before this option, we would have to spend 25 minutes teaching someone how to carefully include only conceptually valuable documents to train the index. With this feature, though, Relativity Analytics does the perfecting itself with zero human intervention. Removing that opportunity for human error and ensuring Analytics can build a good index on its own have helped make it a more comfortable option for folks who are new to the technology.


Given how the industry is evolving, what do you see in the future of Relativity Analytics?

It wasn’t that long ago that people took ESI, TIFFed it, printed it, and sent paper back to the other side—who then scanned and OCRed it. They just couldn’t get past the idea that it would need to be printed at some point. It’s not that people thought this was efficient or optimal—it was just people being creatures of habit. It’s the same way with people who haven’t used analytics yet, and are resistant to embrace it. But the legal industry is catching on, and there’s tremendous opportunity to increase the volume of analytics use.

When it comes to innovation, the closer to the source you can get with analytics, the more opportunities you see upstream for better ROI downstream. We’re seeing interest from practitioners in marching further left on the EDRM with analytics, and we are excited to stay close to these experts as we look for ways that developing technology can open new possibilities and capabilities in the software.

How can e-discovery practitioners encourage adoption of analytics functionality?

This is something analytics engineers have been obsessed with since the beginning. How do you win people over to use a technology that’s ultimately really disruptive? It requires people to change their thinking. We can make things more intuitive, easier to set up, faster to run—giving users an ever-increasing sense of comfort. We can deliver products that clearly show we understand the problems they’re facing. 

For those in the field, there’s education. That comes down to figuring out what’s important to your case team and helping your team and clients understand how analytics can get them there. Check out this blog post or our white papers for tips and resources that can help you do just that.

I like to set it up like this: We’re going to climb a mountain and it’s going to be great when we get to the top—just imagine those views. But there will be muddy streams and slippery slopes along the way, and I don’t want you to be surprised. So here’s an idea of what we’ll encounter that’s uncomfortable along the way. Just know that lots of people have done this before, and no one has ever looked back and regretted the climb.

What do you see in the future of text analytics for e-discovery? Let us know in the comments.

Jeff Gilles joined the kCura team in 2016, where he helps guide development of Relativity Analytics, following eight years of experience developing advanced text analytics technology. He started his career in the software space in 1999, after earning a Master’s degree in mathematics and computer science.