Thought Leaders in Big Data: Interview with Bob Tennant, CEO of Recommind (Part 3)

Sramana Mitra: Could we talk about a few more pain points?

Bob Tennant: We do something called e-discovery, where in the process of responding to a government regulatory request or litigation, you are obligated to provide all pertinent information to that request. To do that, you need to find and identify, then analyze, what information has to be sent over and send it. That process can be inefficient and can be made a lot more efficient with sophisticated technology as well as the capability of doing analysis as we do, with machine learning on top of it.

SM: Where do you apply your e-discovery capability? E-discovery is an industry that has been around for many years.

BT: Like most markets in the technology world, enterprise software has been consolidating for many years. The initial set of problems that people solved in enterprise software were relatively easy to solve. By the mid-90s, most major function organizations had been computerized to a certain extent. From a business process standpoint, most of the wave of technology and the function had been put in place. Since that point, the only major opportunity that I have seen are things that have come as a result of that computerization – hence why storage is a big area topic, why virtualization is a big area right now and why e-discovery is a big area.

Although historically there have been ways of getting through data in a very linear fashion, that process becomes very painful as data volumes rise. Litigation used to be five or six bytes – it was a document on somebody’s desk. Now litigation is five or six terabytes of data contained on many hard drives. Therefore, the process of going through that material needs to be much more efficient than it is right now, and that is one of the ways we get in. We apply techniques to help make that process a lot easier.

SM: So you have compliance, data management, and e-discovery. Where else do you apply your platform?

BT: Another area that can be related is investigation or ad hoc querying of data. In that context you can think about what search has been done for you historically – a search is a kind of ad hoc query, but it is one in which the return is just a long list of documents from which you can do your own research. When you first pose the query, what you are looking for is the answer to a question. What search gives you is not the answer to a question. You will have to find the answer yourself. What you can do is you apply more sophisticated techniques, like a parameterized search where you can drill in. In our case we employ lots of sophisticated machine learning techniques to create aggregates that allow you to visualize or otherwise answer a question. That can be used for investigative purposes in the context of finding bad behavior, understanding what kind of intellectual property you have or a competitor might have, and a wide variety of other uses cases.

SM: Is this an IT sale? Are you selling to the CIO’s office?

BT: It can be both. There are certain investigative use cases where you are dealing with legal teams, and there are certain investigative use cases where you are dealing with business users within organizations. There are also use cases where the CIO would be the person most responsible.

This segment is part 3 in the series : Thought Leaders in Big Data: Interview with Bob Tennant, CEO of Recommind
1 2 3 4 5