DealCockpit™ MCP: Your deal. Your AI. Your rules.

AI in a transactional data room: a topic that merits careful consideration

AI is popping up everywhere. In productivity tools, in CRMs, in accounting software... In data rooms too, teams managing complex transactions are starting to look for concrete answers to a simple question: can AI really help me with a deal, or is it just hype?

We have an answer. And we’ve decided to put it in black and white.

Our choice: letting you choose

An M&A lawyer who has been working with Claude for two years has good reason to trust it. A private equity fund whose IT department has approved Azure OpenAI has good reason to stick with it. A team deploying its own in-house agents has built something that belongs to them.

Our belief is simple: the choice is yours. DealCockpit™ doesn’t make it for you.

That’s why we’ve opened DealCockpit™ to AI assistants via MCP (Model Context Protocol), the standard championed by Anthropic and adopted by OpenAI, which allows language models to interact with business applications using natural language. You connect the model you already use — Claude, ChatGPT, Gemini, Mistral, Perplexity, Copilot, or your own in-house agent — and ask it, in natural language, to act within your data room.

Building above and beyond the standard

MCP is a protocol. What DealCockpit™ has built on top of it is a layer of control and traceability that transforms this protocol into a professional tool for transactional operations.

The AI sees only what you allow it to see. It acts only where you allow it to act. For any sensitive action – deleting files, modifying groups, inviting users – explicit confirmation is required. No irreversible action takes place without a human in the loop. This is an architectural red line, not a temporary limitation.

Every interaction is logged in your audit trail: who initiated the request, what context was processed... You see what is going to happen before it happens, and what happened afterwards.

Our commitments – black on white

We do not train any AI model on your data. Neither our own (we don’t have one) nor that of a partner. Ever.

We do not make strategic recommendations. The AI in DealCockpit™ is an execution and navigation tool. For advice on opportunities or strategic analysis, you should consult a qualified professional. It’s in your best interests, and it’s our firm belief.

Every use case we integrate delivers measurable value on tasks our clients actually perform: not just impressive demos that don’t stand up to the first real deal.

And the MCP is available on all our plans. At no extra cost. Because an innovation of this nature should benefit everyone.

What this means in practical terms

Setting up a data room takes time. Not because it’s intellectually complicated, but because it’s a mechanical process: creating folders, configuring access rights group by group, reviewing logs, and answering incoming questions one by one.

With the MCP, you simply give the go-ahead. The AI creates a folder structure tailored to your type of deal, configures access rights according to your groups, identifies missing documents before your counterparties see them, summarises activity logs, sorts, delegates and responds to incoming questions. You approve. You move forward.

Your deal. Your AI. Your rules.

We see DealCockpit™ as an infrastructure: clear, traceable and controlled, upon which you orchestrate your own intelligence. Not as an autonomous agent that does the work for you.

The more complex the operation, the more valuable this traceability becomes. The higher the stakes, the more essential human oversight is.

That is the position we have chosen. And we stand by it.