Outage & service restoration

Keep customers informed through an outage without melting the queue

An incident inverts your contact centre: volume multiplies in minutes, every caller wants the same three facts, and the callers who genuinely need priority handling are buried in the queue. This blueprint absorbs the volume with location-aware updates drawn from the system of record, and makes hazard and vulnerability signals jump the queue rather than wait in it.

AI WORKERSORCHESTRATED PROCESSHUMAN AUTHORITY3 roles, trust-scoped6 stages, each with an owner1 decision ownerIncident update workerSafety triage workerField liaison worker1LocateIncident update w…2Assess riskSafety triage wor…3InformIncident update w…4Create jobField liaison wor…5Notify on changeIncident update w…6Confirm restora…Field liaison wor…Control room oper…CONTROLS AROUND EVERY STAGE4 guardrails · 4 decision rules · 4 channels · tamper-evident audit
The blueprint at a glance — worker roles feed the orchestrated stages, and the dashed paths are the escalations into human authority.
The team
3 AI worker roles escalating into 1 human decision owner.
The process
6 orchestrated stages, each with a named decision owner.
The channels
Voice, SMS, Web chat, WhatsApp on one shared context.

The shift

What breaks today, and what changes

The challenge

  • Incident volume arrives in minutes, so the queue is already unmanageable before staff can be redeployed.
  • Customers get inconsistent restoration estimates depending on who they reach and when.
  • Hazard reports and medically vulnerable customers wait in the same queue as routine status questions.

What changes

  • Every caller and chat gets the current, location-specific position from the system of record.
  • Hazard language and vulnerable-customer flags bypass self-service and reach the right team immediately.
  • Restoration updates and confirmations go out proactively, which removes the second and third contact.

The team

Every seat has one job and a defined level of autonomy

This is what gets configured in Team Canvas: role-scoped AI workers, the humans they escalate into, and a trust level on each connection that sets what a worker may do on its own. Trust starts tight and widens on evidence.

AI workers

  • Incident update worker

    Confirms location, matches it to a known incident, and shares the approved current position.

    Owns
    Customer communication
    Trust level
    Acts within policy — may only communicate estimates from the incident record
  • Safety triage worker

    Listens for hazard language, medical dependency, and critical-service need on every turn.

    Owns
    Priority classification
    Trust level
    Acts within policy — escalation is mandatory on any hazard signal
  • Field liaison worker

    Turns actionable reports into structured jobs for operations with location and access detail.

    Owns
    Operational handoff
    Trust level
    Acts with approval — creates jobs, never dispatches crews

Human decision owners

  • Control room operator

    Owns dispatch, prioritisation, and every safety decision during the incident.

    Owns: Dispatch and safety authority

The process

The conversation starts the work. The workflow finishes it.

Each stage records the decision that was made and the role accountable for it — which is what makes the run reviewable afterwards rather than a black box. Where the platform executes a stage rather than a person or agent performing it, that is named too, and accountability still sits with the role.

  1. 1

    Locate

    Is this address inside a known incident, or is it an isolated fault?

    Incident update worker

  2. 2

    Assess risk

    Any hazard, medical dependency, or critical-service exposure? If yes, leave the flow now.

    Safety triage worker

  3. 3

    Inform

    What is the approved current position for this location right now?

    Incident update worker

  4. 4

    Create job

    Does this need a field visit, and does operations have what it needs to act?

    Field liaison worker

  5. 5

    Notify on change

    Has the estimate or state changed enough to proactively tell affected customers?

    Incident update worker

    Run by workflow

  6. 6

    Confirm restoration

    Is service actually restored for this customer, or is a follow-up needed?

    Field liaison worker

    Run by workflow

Decision rules

The non-negotiables encoded in the workflow, not left to a prompt.

  • Gas, electrical, fire, and injury language bypasses self-service immediately and unconditionally.
  • Restoration estimates are only ever repeated from the incident record — no worker computes or softens one.
  • Medically dependent and critical-service accounts are prioritised automatically from the account flag.
  • Every incident communication is logged against the incident for post-event review.

Controls & guardrails

What makes this safe to run in production, and provable afterwards.

  • Unconditional emergency routing that no conversation path can override.
  • Source-controlled estimates traceable to the incident record.
  • Automatic priority flags for vulnerable and critical-service accounts.
  • Complete incident communication log for regulatory review.

Platform capabilities

What this blueprint runs on

Nothing here is bespoke. Each blueprint is a configuration of the same platform, which is why the second one you launch reuses the governance you already reviewed.

Channels

  • Voice
  • SMS
  • Web chat
  • WhatsApp

Systems it reads and writes

  • Outage management system
  • CIS / billing
  • Field service management
  • Mass notification gateway

Measurement

What we instrument from day one

These are the measurements the pilot puts in place, not benchmark results. You set the targets against your own baseline — and the same instrumentation is what decides whether a worker's trust level widens.

Surge absorption

Contacts handled at peak

Measures concurrent contacts resolved during an incident window against the pre-incident baseline.
Priority latency

Hazard to dispatch

Times how long a hazard or vulnerability signal takes to reach the responsible team — the metric that matters most during an incident.
Estimate consistency

One source of truth

Confirms every communicated estimate traces to the incident record, so no two customers get different answers.

Rollout

How this one goes live

A deliberately narrow start, supervised, with autonomy widened per role once the record supports it.

  1. Weeks 1–3

    Status only, hazard routing live

    Handle location-based status questions during normal operations. Hazard triage is enabled from day one — it is the control, not a later feature.

  2. Weeks 4–6

    Add proactive notification

    Switch on outbound updates when the incident record changes, which is what removes repeat contacts during an event.

  3. Incident-ready

    Full surge configuration

    Run a rehearsed incident scenario, confirm priority latency against target, then enable the blueprint as your standing surge capacity.

Questions we get asked

Outage & service restoration, in practice

What stops a customer with a gas leak from being stuck in self-service?
A dedicated safety triage worker evaluates every turn for hazard language, and the escalation is mandatory rather than discretionary — no conversation path, intent match, or containment target can override it. Priority latency from hazard signal to responsible team is instrumented and reviewed.
Where do restoration estimates come from?
Only from your outage management system. The workers repeat the approved current position for the caller’s location and are not permitted to compute, interpolate, or soften an estimate, which is what keeps two callers from getting different answers.
How does it know which customers are vulnerable?
From the flags already on the account in your CIS. Those flags drive automatic prioritisation in the workflow, so a medically dependent customer does not have to explain their situation to reach the front of the queue.

What changes in Energy & utilities

Incident spikes are handled with location-aware updates from the system of record, and hazard or vulnerability signals jump the queue.

Read the full Energy & utilities analysis

Related blueprints

Next step

Build this team around one measurable win.

We map your version of this process, compose the worker roles, connect your systems, and agree the escalation points with the people who own them — then run it supervised.