Escalation to a human

What it does

It makes sure a customer whom the automated channel cannot help reaches someone who can. The route is offered plainly, works the first time it is asked for, and is also triggered by the system when it detects failure, distress or a matter outside its scope. The person who takes over receives the history. The business owner of the channel is accountable; operations provide the staff. It operates on every interaction.

How to implement it

• Show how to reach a person from the start of the conversation.
• Act on a request for a person at once, without further questions.
• Trigger handover on repeated failure, signs of vulnerability, complaints and defined topics.
• Pass the transcript and context so the customer does not start again.
• Staff the human channel for the demand that automation creates.
• Measure time to reach a person, and abandonment during handover.

Evidence it operates

Design: escalation rules and triggers; staffing assumptions. Operation: handover volumes and reasons; waiting times; sampled conversations showing requests honoured; test results from mystery shopping the route.

Where it fits

Within the live channel, as the exit from it. It is the primary control for no route to a human and, with vulnerable customer safeguards, for failure to recognise vulnerable customers. It supports exclusion and inaccessibility by giving people the channel does not serve another way through.

Limits

A route exists only if someone is at the other end, and human teams are often cut when automation arrives. Containment targets work directly against this control, so the two have to be reconciled by whoever sets them. Handover to a person who lacks authority to resolve the matter moves the problem without solving it. Out of hours, the route may lead only to a callback.