This morning, LexisNexis announced its Customer Innovation Lab to rebuild “how legal AI gets shipped.”
Lawyers from firms, corporate legal departments, the judiciary, and academia can walk into LexisNexis’s physical space with their specific pain points, and LexisNexis engineers will build against it on the spot. Prototypes get tested that same day, and anything clearing the company’s quality, accuracy, and security bar graduates into a skill or agent on the Protégé platform. OpenAI and Amazon Web Services sit in on the sessions.
If that model sounds familiar, it’s because that’s how Build-A-Bear works, except instead of cuddly stuffed animals, LexisNexis builds high-end, customized legal AI solutions.
CTO Greg Dickason has a video explaining it:
Historically, a company couldn’t conceive, design, and build a legal tech solution within a matter of hours. But with advances in AI-assisted coding, LexisNexis engineers can get to a working prototype quickly, which allows them to design with the end user in the room. That eliminates one of the grand roadblocks to effective legal tech design — engineers used to work for months on a product only to find out after it ships that it doesn’t exactly meet lawyer expectations. Now the lawyers are right there to provide instant feedback.
“Prototypes in hours and trusted products in weeks” is the tagline. It reflects the pressure rapid AI development and the startup mindset have had on legal AI. LexisNexis has mostly watched AI hallucinations proliferate through the legal industry, comfortable that lawyers would prefer to work with an established, careful provider. And while LexisNexis isn’t sacrificing its reputation for caution — the press release notes that prototypes only advance if they clear the company’s established standards — they must feel that the legal community needs assurances that LexisNexis can deliver speed too.
Despite the announcement arriving today, the lab has been running for weeks now:
Already shipping from the Lab: The Lab has already turned ideation with dozens of customers into working prototypes and designs that have advanced into Lexis+® with Protégé™ capabilities coming soon, including:
LexisNexis-developed legal AI models: The Lab is advancing LexisNexis-developed models purpose-built for legal work, using authoritative legal content and deep legal expertise to improve performance, efficiency, and accuracy across complex legal tasks.
An advanced, customer-managed Protégé Vault for analyzing complex multimedia materials. Legal professionals will be able to securely store and analyze documents, videos, images, audio files, and other matter materials. A lawyer could, for example, analyze a two-hour deposition in minutes with Protégé, pinpoint the exact moments when key facts emerge, and trace findings directly to the timestamps and source materials.
Secure Protégé Workrooms for collaboration across teams. Legal professionals will be able to bring together and share documents, analyses, skills, and AI-assisted work related to a given matter with internal teams, outside counsel, clients, experts, and authorized participants, with precise permissions and access controls, all within the same secure LexisNexis environment.
The Lab’s target list for the future involves building drafting tools that adapt to a firm’s own style, precedents, and playbooks; flagging risk against the client’s standards; carrying context across related documents; and fold in litigation, judicial, and regulatory analytics so the draft knows something about the audience reading it.
I don’t know if LexisNexis expected to be compared to Build-A-Bear when they sent along this press release. But the toymaker is one of the few mall-based mainstays that still makes money. Malls were on the decline before the pandemic, but lockdowns turned many a mall into a ghost town. And yet, after cratering to under $2/share in 2020, Build-A-Bear has emerged from hibernation to a comfy $34/share. It’s still a destination for shoppers, because the marriage of personalization and getting a product on-the-spot still sells. There’s no reason that shouldn’t hold true for legal AI too.
Joe Patrice is a senior editor at Above the Law and co-host of Thinking Like A Lawyer. Feel free to email any tips, questions, or comments. Follow him on Twitter or Bluesky if you’re interested in law, politics, and a healthy dose of college sports news.
The post LexisNexis Bets On Build-A-Bear Model For Legal AI appeared first on Above the Law.