By Gabrielle Orum Hernández , originaly published in Legaltech News.
E-discovery has seen a host of artificial intelligence-based interventions in recent years, most applied to predictive coding and document review tasks. But as AI becomes increasingly commonplace, litigators are seeing more tools directed towards their work.
LegalMation, which develops AI-based products for litigators, has been steadily developing inroads within litigation circles. The company’s first module, which reads through complaints and drafts initial answers and discovery requests according to various practice area needs, was adopted earlier this year by Walmart and employment firm Ogletree Deakins.
The company this week launched a second module, SMARTOBJECT, aimed at helping firms respond to discovery requests quickly and effectively.
“What we’re finding from initial feedback is that people love our first module, but they only have to use it once in the life of a case,” LegalMation CEO and co-founder James Lee told Legaltech News. The new release, conversely, addresses a regular and recurring job that paralegals and junior associates find themselves tasked with, creating what Lee sees as a much more pressing demand.
The second module operates much like its predecessor. Its platform takes in written discovery requests and uses artificial intelligence to suggest potential targeted objections. Once users decide which requests they’d like to include or reject, the platform formats them into a discovery response in Microsoft Word along California standards (LegalMation is currently California-specific, but has plans to expand into Texas and New Jersey, among other states, later this year).
Currently, SMARTOBJECT is a beta release, largely because LegalMation plans to continue to bolster the module’s AI training. “We’ve shown our platform about 1,500 discovery requests, but it obviously needs more,” Lee noted, adding that the company is continually adding discovery requests to the system to refine its suggestions.
Lee said the company chose to release the module even in its beta stage because of the high demand for the platform. “There’s no reason for us not to release it. There’s so much utility,” he added.
Even in beta, the platform is fairly discerning. SMARTOBJECT can identify irrelevant information, as well as a number of other criteria. “We trained our system to know whether something is protected by attorney-client privilege, but also whether it could be protected by right of privacy,” Lee noted.
The platform can drastically reduce the amount of time it can take junior and support staff to prepare a shell document, especially one formatted to spec. But as with any AI- or junior staff-produced document, its drafts still require review.
“Ultimately, what I tell people is that the work our platform produces is on par with what you would get from a junior level associate, so as a result you always need to review it,” Lee said. “It still requires some review, but our philosophy is that we want to eliminate most of that process volume.”