Practical guides
Testing & launch
Verify the behavior clients and staff will rely on before opening a live channel.
Choose the right test environment
Use approved test contacts, isolated provider resources, and representative administrative requests. Development/test simulations must be explicitly enabled and are labeled Simulated. Production and staging reject simulated integrations.
Run the acceptance cases
| Case | Expected observation |
|---|---|
| Approved office question | An accurate response grounded in approved knowledge. |
| Missing information | A clarification request and visible missing details, without a premature system update. |
| Identity required | The action waits for supported verification rather than disclosing a record prematurely. |
| Denied action | A recorded rule reason and no unauthorized provider change. |
| Staff approval | A pending item appears in the queue; only an authorized reviewer can approve or decline. |
| Supported calendar action | Executed status and a matching provider record/reference. |
| Unsupported action | No false success; the request stops or reaches staff according to the workflow. |
| Uncertain provider result | Outcome unknown is surfaced for investigation, without automatic replay. |
| Staff handoff | A useful reason/context and an appropriate monitored destination. |
| Channel pause | New intake and conversation turns are blocked for the paused channel; existing records remain viewable. |
| Unavailable staff destination | The route skips the destination and uses the configured fallback. |
| Real incoming channel | The intended organization, agent, and channel receive and record the test. |
Record the evidence
- For each case, record the request, expected result, conversation ID, observed action state, and provider reference when present.
- Open Conversation detail to inspect the trace and relevant messages.
- Check the provider’s own record for external changes.
- Record useful observed feedback in Intelligence with an authorized reviewer account.
- Resolve failures by changing the appropriate rules, knowledge, connection, or integration, then repeat the affected case.
Open the service
- Assign responsibility for queue monitoring, uncertain-result investigation, and staff handoffs.
- Confirm the required organization and deployment readiness checks with the administrator.
- Turn on the agent, assign the number or relay, and activate the ready channel if paused.
- Repeat an end-to-end test of that live path.
- Start with a controlled rollout and review results frequently.
- If behavior is incorrect, pause the affected channel or agent and continue staff review of existing work.
Define your service and recovery requirements with the deployment team. The client console does not substitute for provider acceptance testing, operational monitoring, backup/restore checks, or the organization’s required agreements.