by Jacque Flaherty
on January 24, 2020
Analytics & Assisted Review
Review & Production
The Relativity Fest keynote revealed that since launching active learning in December 2017, the Relativity community had made predictions on more than 750 million documents. That’s more than 500,000 boxes of documents, which is enough to fill more than 150 semi-trucks—an impressive volume, but what’s more impressive are the stories behind these numbers.
The following day, Relativity’s Elise Tropiano led a panel discussion with a few of the Relativity community’s most effective active learning users—Mallory Acheson, Sean Lynch, Nicholas Cole, and Kate Jansons Johns—to hear their stories.
For anyone who’s tried to sell the idea of using active learning on a case to their colleagues or customers, you know it’s usually a journey—some people just get it, and others need much more convincing. Our panelists for this discussion all reported significant roadblocks along the way.
But today, they’re all using active learning “zealously, when appropriate.” In fact, a few noted it being rare when it doesn’t make sense to use it—and even then, they might still sneak it in as a QC safety net.
So, what’s their secret to getting buy-in?
Most of the panelists have treated it as an internal marketing campaign, starting with education and inserting active learning language into the conversation, and then moving to validation that the technology works faster and more accurately than human reviewers. All of them agreed it was integral to have an analytics evangelist internally to help make it happen.
Tactics included using active learning terminology in their coding layouts, using the technology to prioritize review for their associates, running active learning alongside traditional review on live cases to prove superior results, removing any hard cut-offs for case sizes, and focusing the conversation on how active learning will provide their clients better cost predictability and budgeting, as well as the ability to use their own attorneys, rather than contractors, on cases.
Another slightly more enterprising approach included joining every Slack channel in the firm to proactively propose using active learning on cases that sounded like a good fit.
After the panelists achieved buy-in firm wide, they were able to move on to the good stuff—finding more creative use cases for the technology.
Sean noted that active learning’s flexibility came in handy when a regulator seized five million Bloomberg chats and were up against an absolute deadline, but didn’t have a tool that could do anything with the data.
“We recommended putting the chats into Relativity because we knew active learning is not something you need a perfect document to use. We had the entire thing done in 30 days,” Sean said.
Nicholas agreed that active learning’s flexibility is key, but this time, for second requests. The Foley and Lardner team has done three in the eight months leading up to Relativity Fest.
“Active learning is very nimble in these types of situations where what you start with isn’t necessarily what you’ll end up with,” Nicholas said. “The advantage is being able to steer active learning to do what you want to do. Plus, it indicates very quickly where mistakes or gaps are happening.”
Mallory went further upstream and recognized opportunities for active learning in deposition prep and pre-discovery.
“A lot of times we don’t need precision and recall, we just need a flexible workflow that will make us faster,” Mallory said. “There was one case where we didn’t have a formal e-discovery order, but we already had the client’s data and just wanted to know what we should expect. We loaded three million documents, found the key information, and were able to get the case dismissed. Without the quick, upfront information, we could have settled.”
Notably, Kate had a different perspective on the benefits of technology in her firm. Her team focuses on using technology to give attorneys their time back.
“Burnout is very real in law firms. If a nine-to-five schedule is something we all work toward, I can implement technology like active learning to help attorneys find that time,” Kate said. “Some people will use their found time to work harder, some will go home for dinner with their family, and some will use it to find a mentor.”
Active learning in particular helps their attorneys focus on the content that matters most in their cases, which is much more energizing than spending time tagging unresponsive documents. It’s been a way for their team to not only combat burnout, but develop more effective and strategic attorneys—and also furthers their diversity and inclusion efforts.
When it comes to the future of the technology, this group is sure that artificial intelligence (AI), like active learning, will begin to permeate the e-discovery process to increase efficiency and accuracy across the board. And it will be especially effective in the early stages of discovery, when teams are trying to prepare for the problems they see coming.
Another prediction included the prominence of pre-built active learning models that can be taken from one case to the next to create more portfolio clients and increase the sophistication of their work product, rather than reinventing workflows every time.
It seems like we’re on our way to a tipping point of AI and legal technology. Thirty-five percent of U.S. lawyers report that clients are demanding faster service, and 25 percent say AI’s value lies in conflicts clearance, but plenty of teams still struggle just to get started using the technology in their organization.
No worries, though, because the panelists had some advice for that too:
"If you start a [active learning] project and it’s not going the way you’re expecting, keep trying.
Your model can change in the blink of an eye. Don’t give up.”
– Mallory Acheson, head of legal data analytics at Nelson Mullins Riley and Scarborough
“Think about the way you’re branding active learning. For some people, calling it prioritized review
is more palatable. Learn about your people and how they feel about technology.”
– Kate Jansons Johns, litigation support manager at Nutter McClennen & Fish
“Active learning is easy to use on almost any project. The biggest benefit is it will free you up
to think about strategy and look for holes in your case because it’s not all-consuming.”
– Nicholas Cole, director of litigation support at Foley and Lardner
“Play around with active learning. It’s not difficult to set up and you can do a lot with it.
Just remember your starting point should keep the end result in mind. Ask yourself
what you are trying to get out of it and build a model to deliver that.”
– Sean Lynch, director of review services at Ricoh Canada
Jacque Flaherty is a member of the marketing team at Relativity, focusing on research and insights.
Not Using Active Learning? You're Falling Behind
Latest and Greatest in RelativityOne: Active Learning Review Queue and More
Beyond TAR: Active Learning for Linear Review