Skip to content

Legaltech News: Move Over Humans, The Machines are Here to Investigate

JUNE 27, 2017 from Legaltech News

by Rhys Dipshan

Tasked with finding needles in digital haystacks, modern investigations look to move beyond the limitations of human review with technologies like document review and data analytics.

In the digital age, investigations are more about following trails of bytes than breadcrumbs. But while there is inevitably more information than ever to work with, examining it can pose a slew of challenges. Today’s investigators have to adhere to various e-discovery regulations, for instance, or navigate data privacy laws to collect evidence at home and aboard.

And then there is the question of the data itself. While one may be more likely to get at the truth of a matter with more data at hand, reviewing all that data poses a great challenge.

For Alex Southwell, partner at Gibson, Dunn & Crutcher and one of several speakers at Exiger’s June 26 “How Data Analytics Can Transform Today’s Complex Investigations” discussion, the biggest problem rests not with data or law, but with human reviewers themselves.

“The problem is that many companies are trying to [perform investigations] by using the old tools, by throwing bodies at the problems, and the mismatch between problems and solutions is causing dramatic inefficiencies,” he said.

Also speaking at the event was Wayne Matus, managing director of the investigations group at UBS AG. Matus explained that human reviewers oftentimes fail to find all relevant data or even completely agree on what data should be deemed relevant to an investigation.

“If you take a look at human review and analyze it, you see that people are really not very good at reviewing documents,” he said. “Some people can think of human review as the gold standard, but it’s fairly tarnished gold.”

For Matus and other speakers at the event, the answer to human shortcomings lies in modern technology.

By using data analytics and document review tools, “the likelihood of picking out what is significant” in a data tranche is “greatly enhanced,” Matus said. And not only is using such tools far less expensive than human review, but it’s also more considerate of people’s privacy.

In investigations that include a person’s private emails, for example, Matus noted that human reviewers working without technology would need to read every single document to determine their relevancy. “But if you had a technological tool look at those emails, you are no longer infringing on [that person’s] right to privacy,” he said.

And what’s more, such tools can also find relevant information in places that investigators may not have considered.

Brandon Daniels, managing director and president of Exiger Analytics, recalled a time during a difficult financial investigation when he decided expand the investigation’s scope by collecting all communication information relating to an individual during a specific time period and running it through multiple analytics tools.

“One of the critical emails we found was actually a [relevant] conversation between a third party help desk and the rogue trader that wasn’t in [our initial] data set,” he said. The discovery surprised his team, who would have never found the document had they just focused on performing document review on their initial data set.

Daniels’s experience speaks to an often unrealized truth of using tools during investigations: It’s not just about one single technology. “It is actually many different mechanisms you can bring to many different troubles,” he said.

While such technologies include a variety of platforms and tools, Matus classifies them under two categories. The first is what he calls an “older set of tools” like document review that were “intended and designed to reduce the quantity of data” by pinpointing relevant information. The second type are advanced tools that “measure potential relevancy, detect patterns” and connect data pieces into a larger context.

For Rich Plansky, managing director and global head of investigations at Exiger, such advanced tools proved pivotal in a corruption investigation he oversaw years ago dealing with state officials. To assess whether money affected the behavior of elected officials, Plansky’s team looked at disparate data sets like lobbyist disclosure, income disclosure, legislative record, and campaign contribution files.

Using advanced technology, this team then brought “all these data sets together, which were never intended to be looked at together into one common platform, and that gave us context,” he said. “A campaign contribution had meaning if you could follow it all the way to the disposition of a bill whether it [failed] or passed.”

But it wasn’t just the tools that made the connections. While they were able to find connections in the data, the tools still had to be directed and deployed by a capable team of human investors.

“With tools, we can take the salient pieces of information and get [the truth] out of this big mass of data, Matus said. “But there is really no substitute for the individuals who can see the part that a machine can’t.”

He explained that tools used in investigations follow predefined rules and directions on how to work and connect information. But as any good investigator knows, some situations call for an out-of-the-box approach.

“You can program a machine to think the way you think, but that is not good enough,” he said. “You need to have something that is capable of allowing for creativity.”

 

 

See post from Legaltech News.

The risk landscape is constantly changing. Hear about the latest with Exiger.