All use cases

Prevent / Care

Known-incident broadcast with goodwill

Payments just went down — and every affected customer hears it from you first, with an apology, a live status, and a goodwill credit, before the first complaint lands.

Get aheadSystem triggerIn-appChatPushEmail

What it does

When a known incident hits — the payment provider is down, a carrier has failed, a region has a blackout — every affected customer is about to discover it the hard way, and the support queue is about to take the surge. The agent gets there first: it works out exactly who is affected and tells them what happened, what's being done, and when it will be fixed, with a goodwill gesture where the moment warrants one. Outage research backs the instinct: customers who receive 5 or more proactive contacts during an outage score the experience 210 satisfaction points higher. The complaint wave never forms, and the incident reads as care instead of silence.

How it works

  1. 1
    Trigger. a known incident is declared — by a monitoring alert or by an operator marking it — for payments, delivery, or service availability.
  2. 2
    Decision. the agent derives the affected cohort from live data — postcode, order date, SKU, payment method — rather than blasting everyone; per customer it checks channel consent, picks the right surface, and confirms what goodwill the policy allows. Every message passes the judge before sending.
  3. 3
    Action. users currently in the product see an in-session overlay or chat message; everyone else gets push or email — an apology, a plain-language status, the expected fix time, and the promo credit or make-good where warranted.
  4. 4
    Follow-through. when the incident resolves, a closing update goes out and any still-queued messages are cancelled as stale; replies drop into a conversation that already knows the incident context; the inbound surge that didn't happen is measured against past incidents.
Example
AgentQuick heads-up — card payments on our side are down right now, so if your payment just failed, that's us, not you. Your card has not been charged. We expect a fix within the hour, and we've added a credit to your account for the hassle.
CustomerI tried twice. So I shouldn't keep retrying?
AgentNo need — your basket is saved, and I'll message you the moment payments are back so you can finish in one tap. The credit will apply automatically at checkout.

Configuration

How the agent is wired for this use case.

Triggera known incident is declared — by a monitoring alert or by an operator marking it — for payments, delivery, or service availability
Tools & actions
  • Monitoring/status system · receive incident declarations and resolution events
  • OMS / commerce backend · derive the affected cohort from live data (postcode, order date, SKU, payment method)
  • CRM · read per-customer channel consent and history to pick the right surface
  • Promo engine · issue the goodwill credit or make-good the policy allows
  • Messaging channel · send the apology, plain-language status, expected fix time, and credit; send the closing update on resolution
Autonomyevery broadcast message is judge-gated and cohort-targeted so only affected customers are contacted; goodwill runs unattended only within the bounded goodwill and promo policy and customer history
Channelsin-app · chat · push · email
Escalationreplies drop into a conversation that already knows the incident context and route to a human when they go beyond the broadcast

What you need

The inputs this use case runs on. Your channels stay yours; the agent supplies the judgment.

Signals

incident declarations from monitoring or an operator console; resolution events to trigger the closing update

Data

the affected-cohort keys (postcode, order date, SKU, payment method), channel consent per customer, goodwill and promo policy

Guardrails

judge gating on every broadcast message; cohort targeting so only affected customers are contacted; goodwill bounded by policy and customer history; automatic cancellation of queued sends once the incident resolves; suppression for customers already in an open conversation about it

Metrics it moves

  • contact-ratedown: the incident's inbound surge is absorbed before it starts
  • csatup: hearing it from the brand first, with a remedy attached, beats discovering it alone
  • ticket-deflectionup: a single targeted broadcast replaces the surge of identical tickets

Related use cases

Prevent / Care

Preventive CX — reach out before the user arrives angry

the per-user version of opening first

Prevent / Care

Delivery-trouble proactive notice

the same move on a single order's logistics trouble

Prevent / Care

Flight-delay proactive rebooking

incident response that fixes the plan, not just the message

See it on your own customer journey

Bring one drop-off, one churn cliff, or one silent segment. We will show you what a proactive agent with memory and judgment does with it.

Book a demo