Support

AI agent for returns and exchanges

An AI agent records the order number, issue and requested next step so your team can review a return or complaint without repeating questions.

The customer context

The wrong size arrived.

A wrong size and a damaged item need different follow-up. Telp collects agreed information and explains published policies while the decision stays with your team.

What remains after the conversation

The AI agent opened case R-1234 and delivered the exchange reason, order reference and contact to support. The responsible team decides on the exchange.

Telp assistant Website
  1. Customer

    The wrong size arrived.

  2. Telp

    I'm sorry about that. What is the order number, which size arrived, and what phone number can we use?

  3. Customer

    Order 1234, medium instead of large. I'd like an exchange. Ana Example, +381 60 0000000.

  4. Telp

    Ana, I have opened case R-1234 for an exchange from medium to large and sent it to support. They have your order number and contact details, so you won't need to explain it again.

From a question to the next step

  1. 01

    The customer reports an issue.

  2. 02

    The agent collects details and explains published instructions.

  3. 03

    The team reviews and follows up.

Prepare a return request your team can investigate

The AI agent clarifies what arrived, what the customer expected and what they want to happen next. A wrong size, damage and a refund request are different situations. Initial product and order checks are covered in product and order support.

Provide the current reporting procedure and the details your team needs to check. The agent explains published policies; it does not decide customer rights or exceptions. The broader handoff approach is described in AI customer support.

The customer has no order number

Collect alternative details permitted by your process. A missing number should not become automatic rejection of a complaint.

The customer needs an urgent replacement

Record the desired outcome without promising another shipment. Your team checks the order, availability and conditions before agreeing the next step.

FAQ

Frequently asked questions

Does AI approve a return?

This scenario covers intake and preparation. Your team decides approval and the resulting action.

What does it ask when a customer reports a problem?

The order number, item, problem description and desired next step, according to your procedure. It should not request information unnecessary for the investigation.

Can it promise a refund?

Not from the problem description alone. Approving and executing a refund are separate steps that this intake workflow does not confirm.

Does it assess legal deadlines for complaints?

It can convey your approved instructions, but should not provide an independent legal judgement. Unclear deadlines or exceptions go to the responsible team.

What if the customer reports several issues?

The agent clarifies which items and orders each issue concerns. The goal is a clear report, not automatic merging of different replacement decisions.

Does an employee need to ask for the entire problem again?

The transcript is available for follow-up. An employee can ask for the additional information needed for investigation or a decision.

Want to validate this workflow with your own examples?

Tell us the channel, current process and first result you want to improve. We will suggest a focused rollout path and the cases your team should test.