Google Launches Gemini Enterprise for Legal, Challenging Anthropic and OpenAI

Google Cloud unveiled Gemini Enterprise for Legal, an enterprise AI platform for law firms and legal departments built around firm-specific practice skills and Model Context Protocol connections to external legal systems. The platform integrates with case and matter data from Thomson Reuters, document management systems iManage and NetDocuments, e-discovery platforms RelativityOne and Everlaw, e-signature tool Docusign, and legal AI products Harvey and Legora, alongside a case-law database. Consulting partners Accenture, Deloitte, and KPMG will help with deployment. Google Cloud CEO Thomas Kurian said the platform is designed so responses stay grounded in firm-specific data and primary legal sources rather than being trained on client data, with access controls enforced at the connector level. Early adopters named include Cleary Gottlieb, Freshfields, Weil Gotshal & Manges, and Williams & Connolly.

The launch puts Google in direct competition with Anthropic's Claude for Legal, which debuted May 12 with plug-ins across 12 practice areas and more than 20 system connections, and with OpenAI's broader vertical-enterprise push. Notably, the same firms and legal-tech vendors are not picking sides: Freshfields signed on with Claude in May and is now also working with Gemini, while Harvey, Legora, and Thomson Reuters integrate with both Anthropic's and Google's platforms. Per the AI Market Watch index, our pipeline logged 214 Google-related items in the last 90 days versus 203 in the prior period (name-matched over ingested sources only), consistent with Google's accelerating enterprise-AI cadence across finance and legal workflows this quarter.

For builders and investors, the contest is moving away from raw model quality toward the plumbing around it — connector breadth into incumbent systems of record, audit trails, data-segregation guarantees, and workflow-specific skills for tasks like conflict checks and DSAR responses. Vendors without deep integration into entrenched legal software stacks risk being commoditized underneath firms that are actively multi-sourcing their AI layer rather than standardizing on one.