How Deep Dive Is Changing Litigation for Plaintiffs' Attorneys
Elizabeth Koenig spent 8 years litigating at Milbank on the defense side before joining ILS to focus on plaintiff’s ediscovery. As senior vice president of consulting services at ILS, she leads a team of attorneys who provide ESI protocol consulting and project management for plaintiffs in mass torts and class actions.
ILS has worked on major litigation: Volkswagen diesel scandal, 3M combat earplugs, JUUL vaping, opioid crisis. The firm is exclusively plaintiffs-only, never taking work from defense firms or corporate defendants.
These cases share common challenges: massive document sets that can span decades, complex questions of fact buried in millions of pages, and well-funded corporate defendants with the resources to wage prolonged discovery battles. For plaintiffs' firms, the asymmetry is built into the system.
"We're mission driven," Koenig explained. "We're aligned with our clients in that sense that we are attempting to use the technology to level the playing field for individual plaintiffs, harmed parties who are aggrieved by corporate misdeeds and malfeasance."
That mission shapes how she thinks about technology, particularly EverlawAI Deep Dive, which she believes is changing how plaintiffs handle document-intensive cases.
Deep Dive is Everlaw's AI-powered tool that allows attorneys to ask natural language questions directly to their document set like "What did the CEO know about the defect in 2019?", and receive answers with supporting evidence and citations. For Koenig, it represents a fundamental shift.
"Deep Dive, Everlaw's newest tool, is one of the most exciting developments I've seen in over a decade in this industry," Koenig said.
The Resource Gap Is Real
Plaintiffs' attorneys face a structural disadvantage in discovery. They don't have the same staffing levels as corporate defendants. They can't deploy legions of associates to review millions of documents. The economy doesn’t work that way.
"We are often at a disadvantage in terms of attorneys who can staff our cases. We don't have the same resources that the corporate defendants do," Koenig said. "A tool like Deep Dive allows us to do significantly more work with fewer people and to surface critical documents that can be very difficult to find."
She's watched client’s beta test Deep Dive on a massive case with millions of documents, using it almost exclusively for deposition and trial prep. The results have been striking.
"They've been able to cut their review time significantly, improve their ability to get quickly to the critical documents that they need, and see some things that they haven't seen even though they've been working on this case, in these documents, for over a year," Koenig said.
That last part matters. These weren't fresh eyes finding new things; these were attorneys who had been living in the case documents for months, and Deep Dive surfaced materials critical to their case.
"It's really a game changer," Koenig said.
Why Deep Dive Is Different
Deep Dive works through natural language questions, making it immediately accessible for attorneys at all levels, from senior litigation partners to junior associates.
"It’s very accessible because of the natural language use, so they're a great fit for lawyers who are already asking questions. That's part of what lawyers have been doing for hundreds of years," Koenig said. "We ask questions in deposition, in trial, and so now a tool like Deep Dive allows you to do that same kind of questioning."
"A tool like Deep Dive allows us to do significantly more work with fewer people and to surface critical documents that can be very difficult to find."
The adoption pattern has been different. "What we're seeing is lawyers who might not have wanted to get into a database before and let their associates do it, now they do want to get in. They want to ask their own questions."
She compared it to the jump from paper to electronic platforms, noting that people already use similar AI tools in their personal lives. Generative AI platforms like ChatGPT have made the concept familiar.
Practical Applications Beyond Trial Prep
ILS clients are finding uses for Deep Dive that go beyond the obvious deposition and trial prep scenarios.
One application: quality control on outgoing productions. Plaintiffs produce documents too, and they need to ensure nothing privileged or improperly redacted goes out. But they rarely have the resources for multiple levels of review.
"We're able to use a tool like Deep Dive to do a QC," Koenig said. "Maybe we don't have the resources to do a second and a third-level review the way that you'd have a defense firm having a legion of lawyers who are able to review outgoing productions. With a tool like Deep Dive, we can take our production set and make sure we're catching everything."
Medical records present another significant opportunity for efficiency. Mass tort cases often involve lengthy patient histories; records can run three or four thousand pages, dense with medical terminology. Traditionally, firms hire teams of nurses to review them.
Now, using Everlaw's suite of generative AI tools including Deep Dive, document summarization, and Coding Suggestions, attorneys can process medical records much faster. "They're able to use some of the different suite of GenAI tools that Everlaw has to summarize those, to dive into those in a much quicker way," Koenig noted.
Foreign language documents have become more manageable as well, thanks to Everlaw’s AI Translation functionality. Cases involving Asian or European languages used to require hiring attorneys fluent in those languages—an expense that could price plaintiffs out of effective representation. "Now they're able to have much less of a need for that with the GenAI technology," Koenig said.
The Transparency Question
As plaintiffs' counsel, Koenig thinks about how defendants will use AI tools—and how to hold them accountable to discovery rules.
The frameworks for predictive coding and search terms should apply to generative AI, she argues. "Just because these tools are different, just because there's perhaps more of an ease in iterating on the particular prompts that are used, it does not obviate a producing party's need to be transparent," Koenig said.
"What we're seeing is lawyers who might not have wanted to get into a database before and let their associates do it, now they do want to get in. They want to ask their own questions."
She expects courts will start ordering disclosure of prompts the same way they order disclosure of search terms. The validation techniques used for predictive coding can be applied to generative AI workflows. She's already working on protocols and expects court orders addressing GenAI within the year.
There will be fights about this. Not every case will require transparent AI processes, but in high-stakes litigation, it will happen, she believes.
Why the Partnership Works
ILS partnered with Everlaw in 2022. Since then, it's become a critical part of their technology stack.
"Everlaw has a history of working with plaintiffs, with nonprofits, with government entities, which makes it a great fit for ILS," Koenig said.
She's been impressed by how Everlaw has integrated generative AI seamlessly, but also thoughtfully. Deep Dive remains in beta because Everlaw is being careful. "That is, of course, as a service provider, incredibly important for us because we're also standing behind the tool," Koenig explained.
What Comes Next
Koenig sees this moment as an opportunity to develop new workflows clients haven't considered yet.
"This technology is completely different than most things that we've used before. It's, in my mind, a chance to be creative, to bring a new consultative approach to clients," she said.
For plaintiffs facing well-resourced corporate defendants, that matters. Deep Dive won't solve every problem. But for the first time in Koenig's decade in the industry, she's seeing a tool that plaintiffs' attorneys are excited to use themselves, not just delegate. That changes the equation.
