Automated customer service responses: 10 examples that resolve tickets

An automated customer service response works when it resolves the ticket: it reads the real order, completes the action, and tells the customer what happened. Below are ten examples from an ecommerce queue. Each shows the customer message, a weak template reply, and a strong reply that includes what the system actually did.

The order numbers, amounts, and dates in these examples are made up. The shape is not. Every one of these is a ticket type that lands in a store's inbox daily, and the weak versions are close to what most autoresponders send today. Read the strong versions with one question in mind: what did the system do before it replied?

1. Where is my order

Customer: "Where is my order? I ordered a week ago and haven't heard anything since. Order 4318."

Weak template: "Thank you for reaching out! You can track your order via the tracking link in your shipping confirmation email. Please allow 3 to 5 business days for delivery."

Strong response: "Your order #4318 shipped Tuesday with DHL and is out for delivery today. Expected between 10:00 and 12:00. If it hasn't arrived by 13:00, reply here and I'll open an investigation with the carrier."

Why it works: it answers the actual question with a live delivery window from carrier data and pre-agrees the next step, so there is nothing left to write back about.

2. Return request, eligible

Customer: "The sweater from order 4290 is too small. Can I return it?"

Weak template: "You can find our return policy here: [link]. Returns are accepted within 30 days of delivery."

Strong response: "Yes. Order #4290 was delivered 9 days ago, well inside the 30-day window. I've just emailed you a prepaid return label. Drop the parcel at any service point, and your refund of $54.90 is processed within 3 days of the first carrier scan."

Why it works: eligibility was checked against the real order and the label is already in the customer's inbox. The weak version leaves the whole job still to do.

3. Return request, not eligible

Customer: "I'd like to return the boots from order 3781."

Weak template: "Unfortunately we cannot accept your return. Please review our return policy for more information."

Strong response: "I checked order #3781: it was delivered 48 days ago and our return window is 30 days, so I can't create a label for this one. If the boots have a defect, that's a different route. Send me a photo and I'll look at it under warranty."

Why it works: the no comes with the specific numbers behind it and leaves one legitimate door open, instead of ending in a policy link the customer will argue with.

4. Cancellation, inside the window

Customer: "Please cancel order 4402, I ordered the wrong color."

Weak template: "Your cancellation request has been forwarded to our team. We will get back to you within 24 to 48 hours."

Strong response: "Done. Order #4402 hadn't been picked in the warehouse yet, so I cancelled it and refunded $89.00 to your card. It shows within 3 to 5 business days. Want me to place the same order in the right color?"

Why it works: a cancellation is a race against the warehouse. The weak version loses that race by design. The strong one checked fulfillment status and acted while acting was still possible.

5. Cancellation, too late

Customer: "I need to cancel order 4361."

Weak template: "Unfortunately your order has already been shipped and can no longer be cancelled."

Strong response: "Order #4361 left our warehouse this morning, so I can't stop it anymore. What I did instead: I've emailed you a prepaid return label. Refuse the parcel at the door or send it back unopened, and your refund is queued to go out the moment the carrier scans the return."

Why it works: the cancellation is impossible but the customer's goal, getting their money back, is not. It converts the ticket into a return without making the customer ask a second time.

6. Address change

Customer: "I typed the wrong house number at checkout. It should be Elm Street 12, not 21. Order 4415."

Weak template: "Please contact our support team as soon as possible. Note that address changes are not always possible once an order is placed."

Strong response: "Fixed. Order #4415 hasn't been fulfilled yet, so I updated the shipping address to Elm Street 12. The correct address will print on the label, and you'll see the change on your order confirmation page."

Why it works: the address was changed in the store itself, not noted in a ticket for someone to maybe act on. The customer can verify it immediately.

7. Damaged item

Customer: "The vase from order 4287 arrived shattered. This is unacceptable."

Weak template: "We are sorry to hear that. Please contact our support team with photos of the damage."

Strong response: "Sorry, that should not have arrived like that. Reply with one photo of the vase and the box, I need it for the carrier claim. As soon as it's in, I'll ship a replacement the same day or refund the $34.50, your pick. No need to send the broken one back."

Why it works: the photo request comes with a reason and a promised outcome attached, so it reads as the first step of a solution rather than a hurdle before one.

8. Discount question

Customer: "Do you have a discount code? My cart is $120."

Weak template: "Sign up for our newsletter to receive exclusive offers and be the first to hear about our sales!"

Strong response: "There's no general code running right now. Your cart does qualify for free shipping over $100, which saves you $5.95 at checkout. If this is your first order, WELCOME10 gives you 10% and checks itself against your email automatically."

Why it works: the answer stays inside policy. A system that invents a discount to end the conversation trains every customer to ask before every order, which gets expensive fast.

9. Out of stock

Customer: "When is the medium back in stock? I need it before the 30th."

Weak template: "This item is currently out of stock. Please check back later."

Strong response: "Restock is planned for August 2 based on our incoming purchase order, so it would miss the 30th. I've set a back-in-stock alert on your email so you hear the moment it lands. If the date is the hard part, the large is in stock now and ships today with delivery in 2 days."

Why it works: a real date from inventory data, an alert already set, and an honest admission that the deadline won't be met. "Check back later" guarantees the same ticket next week.

10. Angry customer, handoff to a human

Customer: "Third time I'm writing about this. Order 4118 says delivered but I have NOTHING. Refund me now or I'm starting a chargeback."

Weak template: "We understand your frustration. Your ticket has been escalated to our team. Your reference number is 58201."

Strong response: "You've had to ask three times and that's too many, so I won't send you another status update. I've handed this to Sarah with the full history: order #4118, marked delivered on the 12th, no parcel received, carrier investigation opened yesterday. Your refund of $61.20 is queued for her approval so she only has to confirm it. She replies within the hour."

Why it works: the handoff carries the context and a prepared action. The human starts at the decision, and the customer never hears "can I have your order number?" for the fourth time.

The pattern in the strong answers

Go back through the ten strong responses and every one contains a past-tense verb about the system itself. Label sent. Order cancelled. Address updated. Alert set. Refund queued for approval. The weak versions only contain instructions for the customer or promises about the team. That is the dividing line: an automated response that acts closes tickets, and an automated response that explains creates a second ticket with extra waiting attached.

Notice also what the strong answers do with risk. The refund in example 10 was queued, not sent. Money-touching actions can wait for a one-click human approval while everything else runs automatically. Acting and being reckless are different settings.

Whether your stack can produce answers like these comes down to one question: can it act on live store data, or can it only produce text? That distinction, with a test you can run on any vendor demo, is covered in chatbot vs AI agent. And if you want the strong column as your store's default, Nousu does exactly this for Shopify and WooCommerce: widget, inbox, and an agent that takes the action before it sends the reply. See the ecommerce customer service software page for how the pieces fit.