Google's Gemini Enterprise for Legal now plugs into Harvey, Legora, Docusign and RelativityOne via secure MCP connectors. The integration inherits permissions from iManage, NetDocuments and Everlaw, letting legal teams deploy agents across contract review, regulatory scanning and drafting without reconfiguring access controls.
Key facts
- Gemini Enterprise for Legal integrates with Harvey, Legora, Docusign, RelativityOne, HighQ
- MCP connectors inherit permissions from iManage, NetDocuments, Everlaw
- Reusable legal skills learn firm playbooks, citation rules, house style
- Agents cover contract review, regulatory scanning, DSAR fulfillment, redaction, drafting
- No launch date, pricing, or customer names disclosed as of the announcement
Google's Gemini Enterprise for Legal is expanding its integration surface with direct connections to Workspace, Microsoft 365, Docusign, RelativityOne, HighQ, Harvey and Legora. According to @rohanpaul_ai, the new secure MCP connectors will inherit existing permissions from iManage, NetDocuments and Everlaw, meaning firms don't have to rebuild their access-control matrix when adopting the tool.
Key Takeaways
- Google's Gemini Enterprise for Legal adds MCP connectors to Harvey, Docusign, and RelativityOne, inheriting permissions from iManage and NetDocuments, plus reusable legal skills for drafting and review.
- No pricing or launch date disclosed.
What the connectors actually do
The MCP (Model Context Protocol) connectors are the key architectural shift here. Instead of bolting Gemini onto legal workflows with brittle API integrations, Google is using MCP as the standard transport layer — the same protocol Anthropic open-sourced in late 2024 and which has since become the de facto way AI tools talk to enterprise systems. Per the arXiv preprint 2501.00631, MCP provides a standardized interface for tools to expose capabilities to LLMs, and Google's adoption signals the protocol is winning the enterprise integration battle.
What's notable is the permission inheritance. By tapping into iManage, NetDocuments and Everlaw's existing permission structures, Gemini avoids the classic enterprise AI failure mode: giving the model too much access or forcing IT to duplicate role definitions. The connectors read who-can-see-what from the source systems, then enforce those boundaries at inference time.
Reusable legal skills
Beyond connectors, Google is introducing reusable legal skills — packaged workflows that learn a firm's playbooks, citation rules and house style. These skills feed agents that, [per the announcement], can carry work through contract review, regulatory scanning, legal research, DSAR fulfillment, redaction and drafting.
The "house style" piece is the differentiator. Most legal AI tools handle generic drafting; few adapt to a firm's specific citation format and internal conventions. If the skills actually learn these patterns from firm documents, that's a moat — but Google hasn't detailed how the learning happens or whether it's supervised by attorneys.
Competitive read
This is Google's answer to the Harvey-and-friends ecosystem. Harvey has its own integrations with Relativity and iManage, and Legora (formerly Robin AI) has built deep workflow automation. By making Gemini the connective tissue across both rival legal AI platforms and the document management systems underneath, Google is positioning itself as the neutral layer — a classic platform play.
The MCP angle matters beyond legal. If Google standardizes on MCP across its enterprise products, it validates the protocol as the integration standard for AI agents, putting pressure on vendors still shipping proprietary connectors. Legal is just the first vertical to get the treatment.
The announcement, shared by AI commentator @rohanpaul_ai on X, did not include a launch date, pricing, or a list of early customers. Google has not yet posted an official blog or press release, so details on availability and enterprise rollout remain unconfirmed.
What to watch
Watch for Google's official blog post or Cloud Next announcement with launch dates and pricing. Also track whether Harvey and Legora respond by deepening their own MCP support or position themselves as alternatives to Gemini's connector layer. If Google publishes a technical whitepaper on permission inheritance, that's the signal enterprise adoption is real.
