Skip to main content
Monitoring flags Microsoft 365 authentication delays across three managed customers, and Outlook tickets start arriving. The support agent’s assignment is to establish the shared incident, coordinate a consistent response, and carry every affected case to verified recovery or a specific escalation. Configure the supported monitoring, service-health, and ticketing actions in the worker workflow. Your team handles local network probes and unavailable communication routes. This support loop keeps the diagnosis, customer updates, and follow-up connected throughout the incident.

​
Workflow dependencies

Check each dependency against worker tools and connections. Native Teams approval routing is off. Teams messages and customer email need a separately configured Graph, Outlook, Gmail, or help-desk connection and approval for each connector write; assign a person when that route is unavailable. Connector writes require an independent authorized manager to approve the exact request, including external updates, customer replies, and ticket closure. A workflow stage marked automatic does not bypass that approval.

​
From alarm to resolved cases

1

Correlate the incoming work

Retrieve the monitoring event and candidate tickets through configured connections. Group them by symptoms, customer, service, and timing. Keep unrelated tickets separate instead of treating every mention of email as the same outage.
2

Establish the incident evidence

Check the current Service Health record and its affected services. If customer-network evidence is needed, give an engineer a human task to run the appropriate probes from the affected environment and return the results. Record uncertainty when the upstream provider has not confirmed an incident.
3

Prepare a consistent customer update

Build the response from verified scope: the observed problem, the provider’s incident reference when available, confirmed workarounds, and the next update time. Do not promise that another Microsoft service is unaffected unless current evidence supports it.
4

Review the affected cases and communication

Route the proposed ticket list and response for the configured review. For execution, obtain a separate exact-request approval for each ticketing or status-page connector write from an independent authorized manager; a workflow review does not replace these approvals. Assign any manual Teams, email, or customer-specific communication to an owner.
5

Run the follow-up workflow

Configure a supported scheduled trigger or another explicit follow-up mechanism for rechecking Service Health, monitoring, and open cases. Record the incident identifier and the information the next run needs. A reminder mentioned in a chat is not a configured recurring worker.
6

Verify recovery before closing

Compare the provider update with current customer probes and any required customer validation. Prepare the final response with the recovery evidence. Resolve the covered tickets through the permitted workflow; keep exceptions open and escalate them with the incident history attached.
7

Review what the incident taught you

Capture verified organization-wide response guidance with the right audience. Review missed customers, unnecessary escalations, and the clarity of the agent’s updates before changing its instructions.

​
What success looks like

  • Related tickets share a verified incident record and consistent updates.
  • Someone owns each communication path, including channels the worker cannot reach.
  • Follow-up runs use current evidence rather than replaying an old outage summary.
  • An upstream all-clear does not hide a customer whose service is still failing.

Digital workers

Build the agent’s triggers, service access, human decisions, and outcome checks.

Support loop

Carry a request through investigation, coordinated work, and customer resolution.