metamorworks - stock.adobe.com

Microsoft, LexisNexis take partnership deeper with genAI

The cloud provider and the legal research giant plan to create points of integration within Word, Outlook and Teams, while minimizing the risks that come with LLMs.

Microsoft and LexisNexis are taking their partnership to the next level. The companies on Wednesday revealed that they're developing generative AI integration points across Word, Outlook and Teams.

The global provider of legal information and analytics and the cloud vendor have been collaborating for years and currently have several LexisNexis products integrated with Microsoft.

Lexis+ incorporates Azure OpenAI Service. Lexis Connect is a workflow tool built into Microsoft Teams and Microsoft Outlook integration. It features a conversational AI assistant that uses internal corporate documents to answer legal questions.

The vendors now are advancing their collaboration with plans to incorporate generative AI capabilities into Lexis products and integrate them into Microsoft products. They are also working to build technical capabilities so workflow enhancers such as Lexis+ can be made as plugins for Microsoft Copilot when it becomes generally available in 2024.

Law and generative AI

The collaboration comes amid reports that AI will likely disrupt the law industry because it will eliminate most manual work, like drafting tasks such as creating wills, trust documents or merger agreements. It also comes as many law firms are using Microsoft products such as Teams, Word and Outlook, according to the International Legal Technology Association 2022 Survey.

The collaboration will help LexisNexis build an infrastructure that can support Microsoft Office and Teams products and the creation of additional generative AI capabilities that assist in drafting or workflow collaboration, according to Jeff Pfeifer, chief product officer at LexisNexis.

"Our customers have been leveraging certain AI capabilities within our product suite for many years," he said. "What's new or different in the next class or the next generation of development is the ability to create physical work as a result of interactions with generative AI large language models."

For Katie Gardner, a partner in the licensing, strategic partnering and commercial transactions group at law firm Gunderson Dettmer, this collaboration will help lawyers perform tasks such as drafting basic documents without requiring an advanced skill set. At the same time, customers will benefit, she said.

"Lawyers will be happy not spending time doing that stuff," she said. "Clients will be happier not paying for lawyers to do that stuff."

Many lawyers are excited about generative AI, and the collaboration between Microsoft and LexisNexis is particularly promising because it helps lawyers' workflow within widely used Microsoft products, Gartner analyst Ron Friedmann said.

Low risk environment

Moreover, since LexisNexis provides a database that helps lawyers do research, the risk factor with generative AI is lower than just using pre-packaged ChatGPT.

Lawyers are not necessarily taking proprietary sensitive information and asking ChatGPT or another LLM to help them draft a brief from that. Instead they're using the LexisNexis product and Azure OpenAI, which has been adapted for commercial use.

Lawyers will be happy not spending time doing that stuff. Clients will be happier, not paying for lawyers to do that stuff.
Katie GardnerPartner, Gunderson Dettmer

"We're not in the business of saying zero risk. But when you're dealing with a legal research tool ... you're looking at a very different risk equation," Friedmann said. Risks associated with LLMs include hallucinations or misrepresentation of bias and cybersecurity breaches.

To minimize hallucinations, LexisNexis said it is actively trying to expose the LLM to domain-specific tunings related to the legal market.

"As answers are created for our customers, they're relying on exclusive, really highly accurate information that is only available from LexisNexis," Pfeifer said. "What that means is that the answers that are generated for our customers -- whether they be in a browser-based application, like Lexis+, or a workflow application like Word, Teams and Outlook -- all of those interaction experiences are powered by a large language model that is interacting with our core data set to create answers," he said.

However, no matter the risk level, the nature of LLMs means lawyers will still have to be on guard and double-check that the work produced by the generative AI tool is accurate.

"For a long time, we are going to have to be checking our work, because that is a risk that's very well- known. And anyone who's using any of these systems is put on notice of those risks," Gardner said.

Esther Ajao is a news writer covering artificial intelligence software and systems.

Dig Deeper on AI business strategies