Short answer
What is a governed AI coworker?
A governed AI coworker is an AI teammate that can complete work across business tools under explicit permissions, approval rules, spending limits, and an audit trail. It differs from a chatbot because it can act, and from an unbounded agent because the company defines where that action stops.
From answers to accountable action
A chatbot usually returns text. An AI coworker can research, draft, update a project board, prepare a campaign, or assemble a recurring report. Governance makes those actions attributable and reviewable.
- Role-based access determines who can invoke it.
- Connected tools determine what it can reach.
- Always, Ask, and Never rules determine when it may proceed.
- Spend ceilings put a human decision before budget overruns.
Why the work starts in chat
Teams already coordinate work in Slack or Microsoft Teams. Marek is designed to accept a request there, show what it is doing, pause at approval gates, and return the result to the same conversation.
The current Marek status
Marek is pre-launch. A functional prototype is in production QA, there are no public customers yet, and early access will open to small groups in signup order.
Questions
What teams ask before joining
- Is a governed AI coworker fully autonomous?
- No. It can act independently only inside the permissions and limits the company has defined. Sensitive, irreversible, or over-limit actions can require a human approval.
- Can Marek use our existing tools?
- That is the product direction. Marek is designed to work through company-selected connections and credentials rather than forcing teams into a separate suite.