CONTMAN speaks the language of AI. MCP in every module

Monday, 8:30. Anna from accounting opens a conversation with her AI assistant and types one sentence: "Find the leasing contracts that end this quarter, check their notice terms and start reviewing them." Two minutes later she has a list of nine contracts, each with its end date and its key clause, and tasks for the legal team are already waiting in the workflow. Anna has not clicked a single filter.

This is not a vision for next year. This is how work gets done with CONTMAN, because every CONTMAN module speaks the language of artificial intelligence: the Model Context Protocol.

What MCP is and why everyone is talking about it

The Model Context Protocol (MCP) is an open standard that lets AI assistants work safely with a company's systems: read data, take actions and see matters through to the end. Put simply, it is USB-C for artificial intelligence. One connector that any assistant can plug into.

The standard was created by Anthropic, and today it is developed by the Agentic AI Foundation at the Linux Foundation, with OpenAI, Google and Microsoft on board. Claude, ChatGPT, Microsoft Copilot and Gemini all support it. This is not one vendor's novelty but a common language for the whole AI market.

All of CONTMAN in one conversation

Most document systems add AI "on the side": a separate chat window that makes suggestions but cannot get anything done. CONTMAN has gone further. The assistant works directly in every module:

• CONTMAN Archive. It searches documents by content and description, reads them, summarises them, compares versions, adds files and completes descriptions. "Show me every annex to the contract with company X and tell me what each one changed" is a single sentence.
• CONTMAN Workflow. It starts workflows, takes over and decides tasks, and reminds people of deadlines. It can also design a new workflow from a description in plain words: "Invoices over PLN 10,000 are approved by the manager, then by the board."
• CONTMAN Sign. It gathers the documents for signature, checks that nothing is missing and keeps the signers in the right order. You sign, with one click, and the assistant does everything else.
• CONTMAN HR. It completes personnel files, points out what is missing in parts A to E and prepares the paperwork for an employee who is leaving.
• CONTMAN AI. It extracts data from the content of documents and arranges it in description fields, and the assistant puts it to use straight away.
• CONTMAN DP. During bulk scanning it reports the results, points out exceptions to check and sends documents on their way.
• CONTMAN Mobile. A photo of a document taken with a phone goes to the archive, and the assistant takes the matter further before you are back at the office.

One matter, from phone to signature

Piotr is at a client's office and takes a phone photo of a signed order. CONTMAN Mobile sends it to the archive, CONTMAN DP reads the number, the amount and the counterparty, and CONTMAN Archive files the document in the right category. The assistant notices that the amount is over the threshold, so it creates a task for Ewa in finance in CONTMAN Workflow. Ewa approves it in a conversation with the assistant, and the implementation contract is already waiting in CONTMAN Sign for Tomasz from the board to sign.

One matter, seven modules, no retyping and no "has anyone seen this?" emails. People make the decisions, and the assistant does the rest.

Security without compromise

The first question every IT department asks is: "What can the AI actually see?" In CONTMAN, the answer is simple:

• The AI acts as a specific person, with that person's permissions. Anna's assistant sees only what Anna sees and will not do anything Anna could not do herself. There is no "technical account with access to everything".
• Consent for each area separately. Each person decides where their assistant may act: reading documents, changing descriptions, handling workflow tasks, building the archive's structure.
• Every action in the history. When the assistant changes something, the document's history shows who did it and when, with a note that the assistant did it on that person's behalf.
• Sensitive data behind an extra gate. Personnel files require a separate permission, for the assistant too.
• The administrator holds the keys. It is the company that decides which assistants may connect at all.

For security and compliance teams, this is a fundamental change. The AI is not a hole in the permission system but one more user who follows the same rules.

Why this is a huge advantage

You choose the assistant, not the vendor. Claude, ChatGPT, Copilot, Gemini or your company's own agent: anything that speaks MCP works with CONTMAN straight away. When a better model appears tomorrow, you switch without an implementation project and without rebuilding the system.

You invest in a standard, not an experiment. MCP is developed jointly by the biggest players in the AI market under the wing of the Linux Foundation. Companies that build their processes on this standard today are building on a foundation that nobody will switch off.

You gain time your competitors do not have. While others are still copying data from documents into a chat window, your team hands over whole matters in a single sentence. The difference is counted in hours per week for every person.

You are ready for agents. The next stage of AI is agents that run processes on their own. They work on systems that MCP has opened up. Companies whose documents are locked in old systems will be left with AI next to their documents rather than inside them.

Let's talk

Would you like to see how your team hands whole matters over to an assistant in CONTMAN? Book a demo, and we will show you with examples from your industry.