All use cases

Retain / Save

Damaged-order instant make-good

The order arrived broken — and before the customer finishes explaining, the refund, replacement, and courier pickup are already in motion.

Close the loopNext best thingSystem triggerBehavioral triggerTime-based triggerChatEmailPushE-commerce & Retail

What it does

A damaged delivery is the worst service moment a retailer has: the customer paid, waited, opened the box, and got disappointment. The usual answer is a photo upload, a ticket number, and days in a queue — exactly long enough to lose them. The agent runs the entire make-good in a single conversation: it verifies the order, checks eligibility, processes the refund, reserves the replacement, books the courier, and opens the ticket, then follows up to confirm the replacement landed in good shape. Handled this way, the worst moment becomes one of the most loyalty-building interactions a brand gets.

How it works

  1. 1
    Trigger. the customer reports a damaged delivery in chat (often with a photo), or a carrier damage-exception event arrives on the order.
  2. 2
    Decision. the agent authenticates the customer, looks up the order, and checks make-good eligibility against the returns policy and order history; memory tells it whether this customer has had incidents before and whether a refund, a replacement, or both is the right offer.
  3. 3
    Action. inside one conversational thread it sequences roughly seven live API calls — authentication, order lookup, eligibility check, refund, replacement reservation, ticket creation, courier booking — and presents the customer with a done deal, not a process to follow.
  4. 4
    Follow-through. about 48 hours after the replacement is due, the agent checks back to confirm everything arrived intact; if the customer already confirmed, the stale follow-up is cancelled, and if anything is still wrong, the case escalates to a human with the full transcript attached. Every step writes back to the order record.
Example
AgentI'm sorry the mixer arrived damaged — I can see the cracked housing in your photo. Good news: your order qualifies for an instant replacement. I've reserved one in the warehouse, and the courier who delivers it will collect the damaged unit at the same time. Want me to book it?
CustomerYes please. Do I need to print a return label?
AgentNo label needed — the courier brings everything. The replacement is booked, your case number is in your inbox, and I'll check back once it lands to make sure this one arrives in perfect shape.

Configuration

How the agent is wired for this use case.

Triggera damaged-delivery report in chat (often with a photo), or a carrier damage/exception event on the order.
Tools & actions
  • OMS · authenticate the customer, look up the order and contents, reserve the replacement, write every step back to the order record
  • Knowledge base · check make-good eligibility against the returns and make-good policy
  • Payment provider · process the refund
  • Scheduling system · book the courier for replacement delivery and damaged-unit pickup
  • Ticketing · open the case and attach the transcript on escalation
Autonomythe full make-good runs unattended once eligibility passes — refund, replacement reservation, courier booking — and the ~48h check-back is judge-gated; money-moving refunds run under eligibility and value limits, and ambiguous eligibility needs human confirmation.
Channelschat · email · push
Escalationhands off to a human with the full transcript when eligibility is ambiguous or the customer stays unhappy after the make-good.

What you need

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

Signals

damage report in conversation, carrier damage/exception events, order_delivered for follow-up timing

Data

order details and contents, returns and make-good policy, inventory availability for replacements, customer history

Guardrails

eligibility and value limits on automatic refunds and replacements; judge gating on the follow-up message; escalation to a human when eligibility is ambiguous or the customer stays unhappy; cancellation of follow-ups already made redundant

Metrics it moves

  • csatup: the worst delivery moment is resolved in one sitting instead of a ticket queue
  • ltvup: customers whose damaged order is made good quickly become one of the most loyal cohorts a retailer has
  • time-to-resolutiondown: from days of back-and-forth to a single conversation

Related use cases

Prevent / Care

Returns and refunds, judge-gated end-to-end

the broader returns machinery this flow draws on

Prevent / Care

Delivery-trouble proactive notice

catches the problem before it reaches the doorstep

Prevent / Care

Post-resolution follow-up

the check-back pattern, generalized beyond damage cases

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