New AI Agent can now build your knowledge base, connect channels and invite agents — all by chat Try it now
YundaDesk
PricingBlogChannels
Start freeLog in
Questions?Contact sales
Toolkit

Return-Instructions Templates That Turn Returns Into Three Steps

Practical return-instructions templates for eligibility, return addresses, shipping cost, timing, and refund approval boundaries, built for AI-first support with human backup.

YundaDesk Team 2025-12-11Updated 2026-07-10 6 min read

Returns get painful when a customer asks “How do I return this?” and the answer arrives in fragments. First comes the policy. Then the address. Then a separate note about shipping cost and timing. The customer feels stalled, and the support team spends another thread on a routine case.

A better return instructions template turns the process into three steps: confirm eligibility, explain where to ship the item and who pays shipping, then set expectations for refund or exchange timing. AI can answer standard return questions from the knowledge base. Refund approvals, compensation, and price changes should move to a human.

Step 1: Make Return Eligibility Clear

Customers are not trying to read your whole policy. They want to know whether their order can be returned. Start with eligibility, then ask for the evidence you need.

Scenario Template
Standard return Please confirm that the item is still within the return window and remains unused, with accessories and original packaging included. If it meets these conditions, we can share the return address and next steps.
Used item or missing parts This needs a manual review. Please send the order number, item photos, and delivery date so our team can check the next step.
Custom or final-sale item This item may not follow the standard return policy. Please send your order number and we will confirm the available options based on the order policy.

Do not overpromise. AI can explain policy, collect the order number, and request photos. It should not say “this will definitely be refunded” or “we will refund it right away.” If exceptions exist, leave space for human confirmation.

DATA

Return-Instructions Templates That Turn Returns Into Three Steps: routing funnel from intake to high-risk approval

Enter the workflow1,000 conversations
AI collects required context620 conversations
Route to human review280 conversations
High-risk approval45 conversations
Illustrative calculation showing how 1,000 similar conversations may route; not a fixed outcome

Step 2: Explain Address, Shipping Cost, and Packaging Together

Many return conversations get stuck on practical details: where to ship, how to pack, and who pays. Put those details in one reply.

Once the return is confirmed as eligible, please ship the item back to:
Recipient: {return_name}
Address: {return_address}
Phone: {return_phone}

Please include the order number inside the package or write it on the outer box so the warehouse can identify it. If the return is caused by a quality issue or a wrong item, shipping cost will be handled according to our policy. If it is due to size, color preference, or a change of mind, return shipping is usually paid by the customer. Please keep the tracking number and send it to us after shipping.

Cross-border sellers often use different return points by country, order source, or product category. Keep those rules in the knowledge base. Shipping-cost gray areas should not be decided by AI alone. Let AI collect photos and order details, then hand off for review.

Step 3: Set Timing Expectations

Customers can wait more calmly when they know what happens next. Cover three moments: after shipment, after warehouse receipt, and after refund or exchange processing.

Moment Suggested wording
After the customer ships it Please send us the tracking number after shipping. We will record it in this conversation.
After warehouse receipt After the warehouse receives the package, the item will be inspected before the refund or exchange moves forward.
After processing Refund arrival time may also depend on the payment provider. For exchanges, we will share the new tracking information once the replacement ships.

Avoid “instant” or “guaranteed today” unless your policy can always support it. Refunds, compensation, and price changes are high-risk actions. They need human approval and an audit trail. For the boundary checklist, see AI answers first, humans back up.

A Return-Instructions Template for Your Knowledge Base

Replace the fields in brackets with your real policy before adding this to the knowledge base.

Return process in three steps:

1. Confirm return eligibility
Please first confirm that the item is within {return_window} and meets {condition_required}. If the item has been used, is missing accessories, or is custom or final-sale, please contact a human agent for confirmation.

2. Ship the item back
If the return is eligible, please ship the item to:
Recipient: {return_name}
Address: {return_address}
Phone: {return_phone}

Please include the order number inside the package or write it on the outer box. Keep the tracking number and send it to us after shipping.

3. Wait for inspection and processing
After the warehouse receives the package, the item will be inspected. If it meets the policy, we will continue with the refund or exchange. Refunds, compensation, and price changes require human approval. AI does not execute them automatically.

This template is plain on purpose. It gives AI one clean source for eligibility, address, shipping, timing, and approval boundaries. The customer gets a clear next step instead of a long policy excerpt.

Let AI Answer, But Do Not Let It Refund

Returns are a good fit for AI-first support because the questions repeat: Can I return it? Where do I send it? Who pays shipping? How long will it take? Do I need to share tracking? But repeated does not always mean low risk.

Risk level Typical questions Handling
Low risk Eligibility, address, packaging, standard timing AI answers from the knowledge base
Medium risk Shipping-cost disputes, unclear item condition, missing evidence AI collects details, then hands off for review
High risk Refunds, compensation, price changes, complaint escalation Immediate human handoff with approval and audit trail

YundaDesk does not ask AI to make risk decisions for the owner. It lets AI absorb the repeated work before the decision: reading policy, asking for order numbers, collecting photos, and summarizing context. When money or responsibility is involved, the human agent starts with the materials ready.

Keep Return Instructions Consistent Across Channels

Customers may ask about a return in the website widget, send photos on WhatsApp, follow up on Instagram, then add the order number by email. If those messages live in separate tools, the case breaks apart.

Return instructions should stay consistent across channels. Website widgets can show the full three-step explanation. WhatsApp, LINE, and Telegram work better with short replies. Email is useful for the full address and timing. Instagram, TikTok, and Messenger should collect the order number first, then move into the same process.

YundaDesk brings website widgets, custom API, email, WhatsApp, Telegram, Messenger, Instagram, TikTok, LINE, WeChat, VKontakte, Zalo, and YouTube into one workspace with one customer profile. For more, read what an omnichannel inbox actually solves.

Review Return Conversations So the Next One Is Shorter

After the template goes live, review conversations where AI could not answer, the customer asked again, or an agent corrected the reply. Look for patterns: unclear eligibility, wrong return address by country or warehouse, missing shipping-cost exceptions, repeated refund timing questions, or issues that should trigger handoff earlier.

In YundaDesk, when AI cannot answer, when an agent adds the missing answer, or when an agent corrects AI, the system creates a learning suggestion for review. The owner confirms it before it becomes a skill, knowledge entry, or customer memory. Each change is traceable, testable, and revertible. Return instructions can get smarter over time without silently changing themselves.


Good return instructions do not make the policy longer. They make the customer’s next step shorter: whether the item is eligible, where it should go, and when processing happens. Let AI answer standard questions first, and let humans back up high-risk refund decisions.

Run this playbook in your own workspace

AI answers first, humans back up, every step is revertible — everything in this article can be put into practice in YundaDesk.