We have added Model Context Protocol (MCP) servers to ELM, so the AI assistants your teams already use can securely work with your contracts, invoices, and requests.

A year ago, AI in legal mostly meant a chatbot tucked into a sidebar. It was fine for drafting an email and summarising documents, but not much use when you needed it to actually do something. That has shifted. The conversation now is about agents: AI that can pull the right contract, check an obligation, find an invoice, and report back, inside the tools you already work in. What makes this possible is a piece of plumbing called the MCP, an open standard that some people describe as the USB-C for AI. ELM is now part of that ecosystem.

The standard the industry settled on

MCP started as an Anthropic project in late 2024. In about a year, it went from a new idea to something close to a default. The premise is straightforward: rather than building a separate integration for every AI model, you expose your product once through MCP, and any compliant assistant can use it. The old ‘every tool times every model’ problem collapses into a single open interface.

Adoption moved fast. PayPal and Stripe shipped MCP servers, so agents can generate invoices and query payments in plain language. Block, Stripe, PayPal, Anthropic, and Cloudflare worked together on a wave of remote servers. Salesforce now runs a hosted MCP server for Agentforce, Microsoft made MCP generally available in Copilot Studio, and OpenAI and Google added support in ChatGPT and Gemini.

Then, in December 2025, Anthropic donated MCP to the Agentic AI Foundation, a new fund under the Linux Foundation, co-founded with Block and OpenAI and backed by Google, Microsoft, AWS, and others. That was the clearest signal, yet that MCP had become shared infrastructure for the whole industry. There are now more than ten thousand public MCP servers in the wild.

What we built into ELM

ELM now ships live MCP servers across three of its most-used areas, Contracts, Spend, and Requests, exposing almost 30 read-only capabilities between them. In practice, an AI assistant can work on your behalf, strictly within your permissions:

• Contracts — Ask your AI assistant to locate contracts, generate summaries, trace obligations, review amendments, and retrieve audit histories.
• Invoices — Retrieve invoice details, analyse line items, monitor approval status and generate spend reports.
• Requests — Search, open, and summarise requests across your legal team.

Every call runs through your existing ELM login over OAuth 2.0 and SSO, so there are no new credentials to manage, and each one is scoped to exactly what you are allowed to see. Access remains governed by ELM’s existing security model, every action is logged, and the tools are read-only by design. The AI can read and analyse your legal data. It cannot quietly change it.

The point is a simple one. You should not have to go to your legal system to get an answer. The answer should come to you, inside the assistant you are already using.

Why it matters

For legal and operations teams, this is the difference between knowing the answer is in there somewhere and just asking for it. Analysts already credit AI with cutting contract review and cycle times, and the direction in 2026 is toward augmentation: technology that sharpens expert judgement rather than standing in for it. By meeting people inside Claude, Copilot, and the other assistants they have already adopted, ELM turns your legal data from something you log in to retrieve into something that shows up in the flow of work.

Where this goes next

The first release focuses on securely exposing ELM data to external AI assistants such as Claude and Copilot. It also lays the foundation for deeper AI-driven automation across the broader Elevate platform.

We are pursuing certification with the major AI platforms so ELM’s servers can be published as verified integrations, and a Search server plus advanced filtering is next on the roadmap. The goal is an open, secure, permission-aware bridge between your legal system of record and the agentic tools that are reshaping how work gets done.

Want to see ELM’s MCP capabilities in action? Talk to your Elevate team, or contact us.