AL Interview: eBrevia Co-Founder, Ned Gannon, Legal AI Pioneer

Artificial Lawyer recently caught up with Ned Gannon, CEO and co-founder of AI-driven document review company, eBrevia. We discussed how the company got started and how its part in the growing market for legal AI is evolving.


Ned Gannon begins by saying that he was a corporate lawyer before helping to found US-based AI company eBrevia. Earlier he had worked at law firms such as Paul Hastings and the former Dewey & LeBoeuf.

‘I felt the pain. I did a lot of deals where I was sitting there at three in the morning with a cold pizza on my desk because I was manually reviewing documents,’ Gannon recalls.

But, something better was in store. After first helping to co-found Audible Auto, a start-up focused on the intersection of smartphones, vehicles, and social networks in 2009, Gannon took a leap into legal AI, helping to found eBrevia in 2011.

Initially working hand in hand with Columbia University’s Data Science Institute to develop natural language processing (NLP) technology for unstructured data extraction, eBrevia spun out of the academic world a year later in 2012. Although, the University still owns equity in the business.

Back then they joined a small, but pioneering group of legal AI companies that had not yet hit the headlines, but were busy improving their NLP and machine learning software to bring AI to the legal market.

The company, while a little different to peer companies such as Kira and RAVN, is fundamentally offering clients the same deal: AI-driven review of their documents and the rapid, more accurate extraction of provisions that the clients are looking for.

screen-shot-2017-01-27-at-08-03-31
Ned Gannon, CEO, eBrevia

But, what now? In 2016 it seemed that legal AI had come of age at last and that law firms had finally embraced its potential. Where is eBrevia in all this and how does Gannon see the future unfolding from 2017 onwards?

In the fall, eBrevia received an investment from one of its strategic partners Donnelley Financial Solutions. Gannon says that with the backing of a large public company that’s a trusted name in the legal and finance industries, eBrevia is able to easily scale to serve the needs of their multinational clients. The Donnelley Financial Solutions salesforce is selling eBrevia’s products globally.

According to Gannon, eBrevia’s clients include Fortune 500 companies, audit/consulting firms and about 15 law firms including global law firms. Medium-size US law firms such Reitler, Bailey Duquette and Day Pitney have also leveraged the technology.

As to the future, Gannon is bullish. ‘This is an increasing market space and we’re drinking from a fire hose. We’re just going to keep on pushing forward,’ he says.

Gannon adds that the offering has developed also, with a more DIY facility now on offer that allows clients to custom build their own search provisions, something that law firms often say they want now, rather than relying on pre-sets, or needing to buy the time of the AI company’s staff to create what they want.

This move also reflects how eBrevia, if not the whole legal AI market, has evolved in recent years. Initially when the focus was very much on M&A due diligence document review, the company had employed corporate lawyers to help build search models for the NLP software, or as other companies call them ‘recipes’.

As Gannon explains: ‘Once we’d done this for M&A, we then did the same for real estate and other areas. It was an intermediate step. Now we’re putting the power totally in the clients’ hands.’

That is to say, as the knowledge of the legal AI companies has expanded and as the recipes have been refined, the power to control what is searched for in terms of legal clauses and other unstructured data is increasingly handed over to the clients to do what they want in a more bespoke way. To some degree one could say this has been a learning curve for both the providers of legal AI document review and the buyers of it.

One growth challenge that Gannon, as well as many other experts in the legal AI world, recognise is that selling into inhouse legal teams is not the same as selling into a law firm.

‘In large organisations there can be a variety of gatekeepers and this can slow things down. Some inhouse lawyers have been very innovative, others are more sceptical. Even with certain industries you see no pattern,’ Gannon observes.

That said, Gannon notes that companies in the energy sector are increasingly looking to use AI in their inhouse functions. The exact cause is not clear, but perhaps one reason is that the energy sector is a tight knit community and ideas travel fast between companies.

And lastly, as the wider legal community becomes more familiar with AI, what will happen in terms of the different AI companies competing? How will clients decide to use one company rather than another?

Gannon accepts that it is a challenge to compare different companies’ NLP software. While accuracy and an intuitive user interface are important, Gannon sees customers’ evaluation of providers going beyond just the software itself.

Clients want to work with a company that is going to continue to innovate as the field evolves and that can service their global needs. Gannon notes that: ‘Our sales and project staff were lawyers so they understand the issues clients face. Similarly, our relationships with Donnelley Financial Solutions and Columbia University are an important part of our value proposition.’

And pricing? Will we see a price war? Gannon is unsure. At present it would appear that most AI companies are charging about the same rate per document reviewed, although some have different charging models related to volume or data size.

The more likely differentiator, Gannon notes, is that there will many more developments to come in legal AI. ‘AI is going to get more sophisticated and how far it can go in terms of taking on more complex legal tasks is a matter of time,’ Gannon says.

eBrevia and other peers of the first wave of legal AI companies that started in the early 2010s have already taken things a long way. eBrevia, for example has found that its accuracy in document review is at least 10% higher than using human lawyers alone, while in terms of overall time for review there is a reduction of 30% to 90%, with the ability of the software to extract data from over 50 multi-page documents in less than a minute.

But, as Gannon says, they have reached this level in just five years of working on NLP and machine learning. What happens next is anyone’s guess, but Gannon is interested to see how smart contracts and AI will link together in the years to come, although smart contracts are still very much in experimental mode for now. There will also likely be increasing capabilities in terms of what the AI review software can do, moving up the value chain inside a law firm or corporate legal department.

To conclude, although eBrevia has been in the legal AI market since it started over five years ago, they clearly have everything to play for now the sector has come of age. And as Gannon notes, things are only just getting started in terms of where this will all go.