Bundledocs- Why model context protocol matters for the next stage of legal AI adoption
The conversation about our use of AI in the workplace is moving at a rapid pace. We go from discussing what AI is and what it can do, to how it connects with systems and processes in the blink of an eye.
Its use in professional services is accelerating. Thomson Reuters’ 2026 AI in Professional Services Report found that genAI is now well embedded across legal, tax, accounting, risk and government sectors, with growing interest in how agentic AI may reshape professional work.
However, when it comes to the legal profession, there is still a difference between access to an AI tool and making AI useful. Legal work rarely happens in a standalone environment. Documents sit within document management systems, bundling tools or shared workspaces. They move between support teams, counsel, experts and clients. They also carry permissions, version histories, review stages and final approvals.
For AI to make a real difference in the legal world, it needs to connect with those systems safely and practically. That is where model context protocol (MCP) comes into play.
What MCP means in practice
In simple terms, MCP is an open standard that allows your AI applications to connect with external systems, tools and sources of information in a more consistent way.
Until now, much of the focus around genAI has been on the AI tool itself. A user asks a question, provides information and receives an answer. That can be useful, but it often relies on people manually moving content between systems.
MCP is different and more connected. With the right permissions and integrations in place, compatible AI assistants can interact with the systems where legal work is happening. In practice, that could mean helping teams transfer, update or prepare documents as part of connected bundling, binding or review processes, rather than requiring users to move information manually between separate platforms.
For law firms, the value is not in the technical standard itself. It is in what that standard may make possible.
Why legal work needs connected AI
Legal work is document-heavy, process-driven and highly controlled. A single matter can involve multiple contributors, document versions and access levels, making standalone AI difficult to scale.
If an AI tool sits outside the systems legal teams already use, documents may need to be exported, prompts prepared, outputs checked and content moved back into the right place. Instead of reducing friction, AI can risk adding more steps.
This article originally appeared in LPM. To read the full article, click here.
