AI asistenti a MCP

The connection through which your own AI assistant reads and fills in the walkthrough, as you. What it does, what it may not do and how to start.

An AI assistant collects, sorts and enters facts. Your company carries out and approves the measures. What NIS 2 says about AI, and where the line runs.

Art. 20 NIS 2 puts approval and oversight on management, whoever did the typing. Whose entry it is in the walkthrough, and why no assistant approves.

In 19 of 21 items, item by item: fields, lists, ratings, agreements, sign-ins, backups, policies. Finishing and approving stay in the app.

An admin turns assistants on in Settings; until then they are off. Where the switch is, who may flip it and which accounts can connect after that.

One command, one browser sign-in, and Claude Code reads and fills your walkthrough. With a key for scripts and a first task for 5.1.

As a custom connector with “Register automatically”. Which plan, who adds it, and why Claude needs an instance reachable from the internet.

Add the walkthrough as a tool of an agent, with OAuth 2.0 and Dynamic discovery. Why OAuth rather than an API key, and what the agent may do then.

Browser sign-in or key: when to use which, how long tokens and keys last, and what ends a connection.

That depends on the server. What the walkthrough's MCP server checks on every call, what an assistant cannot do there and what stays with you.

What the assistant reads goes to its vendor, under your agreement with them. What that means for the DPA and training, and how to keep the data in house.

As a custom MCP server with a browser sign-in, since ChatGPT takes no pasted key. What has to be reachable for it.

What GRC vendors' MCP servers may read, write and delete, who switches them on and what stays with a person. As of October 2026.

What an assistant can keep up in an ISMS, what top management and the certification body keep, and where the NIS 2 walkthrough does groundwork.