Welcome

Deloitte Alliance Illustrates Growing Role for Data Extraction Software

March 8, 2016, 5:57 PM

Kira Systems announced a new partnership on Tuesday with Deloitte U.S. to introduce more corporations to its software — which uses machine learning to automatically search large sets of unstructured data and extract specific terms.

Noah Waisberg, CEO and co-founder of Toronto-based Kira, said Deloitte is the company’s biggest client and has over 3,000 active users of its extraction software, mainly in the auditing and consulting businesses. The software has a number of use cases including due diligence reviews, and any situation where a large data set needs to be searched, Waisberg said.

Under the new partnership, Deloitte will continue using Kira, but it will also try and sell Kira to its clients for ongoing use.

“They think there’s a lot of opportunity to sell this on to their clients in conjunction with services,” said Waisberg, who added that he believes Deloitte’s use of Kira represents the largest deployment of machine learning and artificial intelligence at a professional services firm.

Craig Muraskin, managing director of the Deloitte’s U.S. Innovation Group, would not disclose the financial terms of the agreement with Kira. Deloitte could refer some of its clients who want an independent agreement with Kira, or its clients may choose to sub-license Kira through Deloitte, Muraskin said.

Waisberg also declined to disclose details.

“They’re not getting an ownership stake in the company,” he said. “In some cases, there will be revenue-sharing [if they were referring us to a client]. In other cases, they will pay us to use our software.”

Waisberg said that 20 to 25 percent of all U.S. companies have a contract management system, which is either a database or an unstructured storage system for all of a companies’ contracts. One possible scenario for Deloitte would be to help clients create a contract management system, and then sell Kira so the client can easily search their contract management system, he said.

Said Muraskin: “We continue to see ever more opportunity to leverage the technology.”

He said Kira can extract and search terms from “pretty much any unstructured data,” including 10-Ks, 10-Qs, board minute meetings, invoices, transcripts to calls and a variety of other formats. This can be helpful during an audit when Deloitte is trying to ensure that a company’s disclosures match reality because it allows the company’s professionals to search a wider data set.

The system is also easily programmed to search new formats, Muraskin said.

“We’re using it on tweets, for example,” he explained. “We’ll say, ‘Show us all instances of X,’ when people are saying something about a product, positive or negative things. This makes for a much more efficient reading of the data.”

Waisberg said Kira is also useful during transactional work in figuring out overlapping licenses: “When you read a [merger] announcement and they say there will be $300 million in synergies, some of the people at Deloitte are the ones figuring out that you have a license to Microsoft Office and we have a license, but you have a better price, so we’ll keep your license.”

Amy Harvey, practice innovations analyst at Chapman and Cutler, said her firm started using Kira in January 2015, and several practice groups have developed client-specific and general uses of the software, such as deal analysis.

But Harvey said she also uses it for internal knowledge management because she can train Kira to pick up language within attorney-work product, and search for specific meta data across the entire firm.

Founded in 2011 by Waisberg, a former Weil Gotshal & Manges associate, the company was originally called Diligence Engine and planned to target law firms as its main client. Waisberg said his experience in Big Law had taught him that junior lawyers were working on matters, such as due diligence, that would be more efficient if there were technological assistance.

The software, which uses machine learning and artificial intelligence algorithms, was only supposed to take a few months to build, but ended up taking two-and-a-half years before it even worked at all, he said. Kira has never taken any outside investment, and so Waisberg said it was financed out of his own and his co-founder’s personal savings.

Today, the company still has a dozen global law firms as clients, he said, but two-thirds of its clientele are corporate law departments and professional service organizations such as Deloitte. It has also grown from four employees to 25, according to Waisberg.

“When we first started selling to law firms, it was really tricky because they didn’t know our software existed as a category,” he said.

It wound up changing the name because extraction software is being used for more than simply due diligence reviews, Waisberg said. A number of competitors that produce extraction software have also launched including eBrevia, Seal Software, Ravn, Recommind and others.

Deloitte, for instance, has a U.K. partnership with Recommind, which makes a tool called Perceptiv that extract terms from derivative contracts. Together, the partnership provides a managed service for banks’ Over-the-Counter International Swaps and Derivatives Association agreements.

In terms of the growing competition in the field, Waisberg said: “When we started doing this, there literally were next to none. And now there are a lot of companies doing this.”

(UPDATED: This post has been amended with the correct spelling of eBrevia.)

To read more articles log in.

Learn more about a Bloomberg Law subscription.