Docs / Start here

What Mindset does not do

Worth reading before the rest, because most disappointment with a platform of this kind comes from an assumption nobody checked.

Mindset does not find agents you do not know about

If somebody builds an agent in Claude, ChatGPT or Copilot Studio and connects it straight to a system, Mindset never sees it. Discovery belongs to your tenant admin tooling, your data loss prevention and your threat detection platform.

Discovery and Mindset work together. Discovery tells you something exists. Mindset is the record of what was allowed. Anything discovered and absent from Mindset reached its system by some other route, and that absence is the finding.

Mindset does not drive a browser

There is no browser automation, deliberately. An automation that works by clicking through a screen leaves no record anyone can check afterwards. If the only way to reach a system is through its interface, that automation stays where it is until the system has an API, a database, a sheet or an MCP server.

A run does not resume

If a run stops part way, it stops. Nothing picks it up from where it got to. Re-running is a decision a person makes, and a re-run is a new run that starts from the beginning. Anything that has to survive a failure and continue belongs in a queue, not here.

Mindset does not replace your identity or access tooling

Who may use which model, who is licensed, who is in which group. That stays where it is today. Mindset governs what a sanctioned agent may do once it exists.

Mindset does not fix an automation nobody had defined

If the work has never been written down as steps, Mindset will not discover them for you. The three-question test for whether something is ready to move is in Setting up Mindset.