Skip to main content
A business customer reports that calls connect but audio works in only one direction. The support agent’s assignment is to carry the case from the ticket to a working call: collect the evidence, coordinate diagnosis and authorized remediation, obtain a meaningful test result, and respond to the customer. Use configured cloud connections for the agent’s ticket and service-system work. An engineer uses Studio for live SBC captures, firewall inspection, and changes that need local access. The support loop makes that handoff explicit and returns its findings to the case.

​
Workflow dependencies

Check the required worker tools and service connections before delegating actions. Access to a desktop terminal, packet capture, or browser session does not grant a cloud worker the same access. 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 complaint to verified resolution

1

Read the ticket and identify the calls

Retrieve the request through an available integration. Establish which party could hear the other, the affected call direction, example timestamps, relevant numbers, and whether the issue is repeatable. Ask for missing details before selecting a trace.
2

Collect authorized evidence

Fetch the matching call records and SIP evidence through the configured service connection, or request them from the responsible engineer. Search eligible organization-visible Org Brain sources for previous guidance. Keep the evidence scoped to the customer and incident.
3

Hand off the live investigation

Give the engineer the symptoms, example call identifiers, and collected evidence. In Studio, the engineer inspects SIP and SDP negotiation, traces RTP in both directions, and checks the SBC and firewall state. Use a controlled test call to distinguish the observed fault from a familiar but unverified diagnosis.
4

Review the remediation plan

Record the proposed device changes, customer impact, validation method, and rollback plan. Route the decision to the configured approver. The agent should not report a fix as applied when it has only prepared instructions or requested approval.
5

Carry out the permitted work

Use the configured execution path for each approved action. An engineer performs local SSH or firewall work in the desktop workspace and returns what changed and any errors. Keep the case open if the change cannot be completed or requires a further decision.
6

Prove the customer outcome

Place a representative test call and verify audio in both directions, including the originally failing call path. Record the time, path, and result. If the test still fails, return the evidence to investigation or escalate to the appropriate specialist.
7

Respond and resolve

Prepare a concise ticket reply with the confirmed cause, completed action, validation result, and any remaining customer step. Send through the permitted ticketing integration or an assigned reviewer. Resolve according to the agreed acceptance criteria rather than the fact that a command ran successfully.
8

Improve the next response

Save the verified configuration lesson with the right audience and update the engineer’s diagnostic procedure where useful. Review the agent’s run for missing questions or weak handoffs before changing its instructions.

​
What success looks like

  • The engineer receives a focused case with the relevant call evidence and a clear question.
  • Infrastructure changes are tied to the case, approval, and validation method.
  • The final reply describes a tested result; a remaining fault becomes an evidence-backed escalation.
  • Reusable knowledge helps the next investigation without replacing live verification.

Digital workers

Connect ticket handling, service tools, human decisions, and outcome review.

Terminal

Use the operator workspace for local SBC and firewall investigation.