Google Automates Legal Work With Gemini Enterprise for Legal

Gemini is both the AI model built by Google DeepMind and the brand name for the app that runs on it. This new release isn't a new model — it's an enterprise product that adds legal-specific connectors and workflows on top of the existing Gemini model.

It Plugs Directly Into Existing Legal Systems

The core of this product is the MCP, or Model Context Protocol, connector — a common standard for linking AI models to outside tools and data, often described as the "USB-C of AI." Through this standard, Gemini Enterprise for Legal connects to document management systems like iManage and NetDocuments, e-signature service DocuSign, e-discovery tools Everlaw and RelativityOne, research platform Thomson Reuters HighQ, and legal-specific AI service Harvey. Because the integration keeps a company's existing access permission structure in place, the rules governing who can see which documents don't change.

Ready-Made Agents From Partners Like Deloitte Come Bundled In

Partners such as Deloitte separately sell ready-made AI agents built for specific tasks, like summarizing contracts. These industry packages essentially boil down to pre-built prompts — called "skills" — combined with the relevant data connections. A skill is a bundle of task procedures that an agent can run and modify; once it's configured to match how a company actually works, nobody has to rewrite instructions from scratch every time. The model underneath is the same Gemini that powers Google's other standard products.

After Finance, Healthcare Is Next as the Industry Lineup Grows

Google Cloud CEO Thomas Kurian said the product is being offered in preview as part of a broader lineup of industry-specific AI solutions. A financial services version launched the same day, with a healthcare and life sciences version expected to follow.

Industry	Status
Legal	Launched in preview (integrates with iManage, DocuSign, Everlaw, Harvey, and more)
Financial services	Launched same day
Healthcare & life sciences	Coming soon

Anthropic Is Already Heading the Same Direction

Google isn't the first to release an AI package for the legal industry. Anthropic already offers plugin-based solutions targeting several industries, including legal. The two companies use different names, but the underlying structure looks similar.

Category	Google (Gemini Enterprise for Legal)	Anthropic
Connection method	MCP connector	Plugin
Underlying model	Same Gemini as the standard product	Same Claude as the standard product
Work unit	Skill (pre-configured prompt)	Skill (pre-configured prompt)

Anthropic has also recently expanded authentication for Claude's MCP connectors, deepening its enterprise integrations. It's a sign that competition among model makers is shifting away from building new models and toward how deeply they can plug into existing workplace systems.

Who Can Use It, and How

Since Gemini Enterprise for Legal is still in preview, it isn't a consumer app — it's an enterprise product that law firms and corporate legal teams adopt through Google Cloud. Organizations that bring it on connect their existing system accounts, like iManage or DocuSign, through MCP connectors, layering AI capabilities on top while leaving their current access permission settings untouched.

Based on what's been described so far, here's what that looks like in practice. Upload a batch of contracts, and it flags risky clauses or unusual language to cut down review time. Ask about a specific case or regulation, and it pulls together the relevant legal research for you. It can also track regulatory changes and flag them as they happen. On top of that, you can add task-specific agents built by partners — like the contract-summarization agent Deloitte sells.

Editor's Take

What stands out here is that there's no new model anywhere in this announcement. Both Google and Anthropic rolled out industry products for legal and finance without touching the model itself. That tells you the competition has shifted from "who builds the smarter model" to "who can plug into existing work systems with the least friction." That's exactly why the list reads like a who's-who of software law firms have relied on for decades — iManage, DocuSign, Everlaw, RelativityOne. For a law firm, adopting new AI shouldn't mean changing systems; the AI has to come to the systems they already use.

Looked at from a practical standpoint, law firms and legal teams in Korea don't need to rush out and adopt a new model right now. The more important questions are whether the document management and e-signature systems already in use support a common standard like MCP, and whether partner firms can build skills tailored to the actual work. Repetitive, well-defined tasks like contract review are where ready-made agents will show results first, while areas that require judgment — like interpreting case law — will likely keep needing a human to sign off, for the time being at least.

In the next few months, Google's industry lineup will likely round out with the healthcare and life sciences version. By then, other model makers will probably have rolled out similar industry-specific packages of their own — and this market will end up being decided not by whose model performs best, but by who manages to lock down the broadest reach across each industry's software ecosystem.