Example walkthrough · Acme
See how a requested screenshot gets approved, run, and recorded.
This is an example, not automatic fixing. A connected AI app asks for a screenshot of an Acme Corp Mac (acme-mac-01). The request waits in Actions until a person approves it. Then it runs, and the record shows the result to whoever picks up the support request (ticket) next.
The path
Four steps from request to record.
- 01
Request
A person, or an AI app connected through MCP, requests a screenshot of an Acme Corp Mac (acme-mac-01). It does not run yet.
- 02
Review
In Actions, a person checks what was asked, which computer, and the risk. Approve & run dispatches immediately; Reject stops the work and keeps the record.
- 03
Execute
The FlowRMM agent takes the approved screenshot on the computer. The image, exit code, and status come back on the same action.
- 04
Record
The record shows requested, approved, sent, and completed, so the next person on the support request is not starting from scratch.
capture_screenshot → acme-mac-01
Same console, live agent
See the machine, then act with accountability.
Operators can open Remote Console when they need eyes on the desktop, then use Actions for commands that deserve a second look. Governance sets four-eyes defaults and when self-elevate is allowed.
Approval gates
Autonomous, with a person where it matters.
FlowRMM is the hands of FlowSuite. When FlowPSA picks up a ticket, AI agents run the fix through FlowRMM on the real computer, verify it worked, and close the ticket with proof attached.
Every action can be gated with human-in-the-loop or four-eyes (two-person) approval, and high-risk actions always wait for a person on Actions before they run.
Next
Try this with your own support requests.
Book a demo to see remote access and approvals, or join the Business waitlist to get set up in your own Azure subscription.