Supervision and escalation

People can step in at any moment.

Pause or cancel any run while it happens. When an agent reaches a step it may not take alone, it hands the decision to a person.

Live run Refund case 4821
Live
  1. Read the refund request
  2. Check the order history
  3. Refund above team limit
Escalated Needs a decision from Priya, Support lead

The agent paused this step. Reminder sent after 2 hours.

Approve Pause run Cancel

Oversight that lasts the whole run.

Not just at the start

Supervisors can pause or cancel a run at any step.

Hard cases go to people

The agent stops and asks instead of guessing.

Nothing gets forgotten

Waiting decisions send reminders and can alert an admin.

Escalation to the right person

When an agent reaches a step outside its limits, it pauses and sends the decision to the right person, with the context attached.

Live run Refund case 4821
Live
  1. Read the refund request
  2. Check the order history
  3. Refund above team limit
Escalated Needs a decision from Priya, Support lead

The agent paused this step. Reminder sent after 2 hours.

Approve Pause run Cancel

Pause or cancel any run

Supervisors watch runs as they happen and can pause or cancel one at any step.

Evidence and artifacts Run trace
Live
  1. 09:41 Request classified
  2. 09:43 Scope checked
  3. 09:48 Evidence gathered
  4. 10:02 Approval waiting
Artifacts Contract delta.pdf Spend summary.csv Recommendation.md

Reminders when work waits

If a decision waits too long, the agent sends reminders and can alert an admin, so nothing stalls quietly.

Workspace
Approval request
APPROVAL GATE

Vendor outreach is blocked

Finance and Legal must approve the packet before any external message can be sent.

How it works

From request to recorded result.

  1. The agent works

    Inside its limits.

  2. It hits a hard case

    And stops at that step.

  3. A person decides

    With the context in front of them.

  4. Work carries on

    Or is paused or cancelled.

Every feature runs inside the same six controls.

Identity, access, approvals, supervision, audit and policy apply to every agent and every run, whichever part of the workspace it happens in.

See governance and audit
  • Identity
  • Access
  • Approvals
  • Supervision
  • Audit
  • Policy
FAQ

Supervision and escalation: questions and answers

Short answers to what teams ask before they put this part of SoftworkerAI to work.

Contact us directly
Can we stop an agent in the middle of a run?

Yes. Any run can be paused or cancelled at any time by the people allowed to supervise it.

Who does an agent escalate to?

The person or role you set for that kind of decision, such as a team lead or a finance approver.

What if nobody answers?

The step stays paused. The agent sends reminders and can alert an admin.

The multiplayer workspace for humans and AI agents is how teams move beyond prompting and safely hand AI real work. It gives people one place to plan, approve, and oversee execution, so they can run work with AI agents they can govern and trust.
The SoftworkerAI Team