--- title: Agentic returns url: "https://parcellab.com/copilot/agentic-returns/" type: Page date_published: "2026-11-14T17:53:20+00:00" date_modified: "2026-11-14T17:54:44+00:00" description: "The parcelLab Returns Agent checks every return against your policy, offers an exchange or store credit first and books the label, in chat, email or WhatsApp." --- # Returns handled by an AI agent, inside your rules The parcelLab Returns Agent checks each request against your policy, offers an exchange or store credit before a refund and books the label, in chat, email, WhatsApp or your returns portal. [Book a demo](https://parcellab.com/book-a-demo-platform-tour/) [Try the demo](#try-it) 81 % of 411 customer emails resolved by the parcelLab WISMO/R Agent with no human touch, in a two-week pilot at a UK retailer 11 % higher customer retention through exchanges and store credit 18 % fewer returns at Conrad after moving to the self-service returns portal ## Run a return through the agent Click through the returns portal and the AI agent on a sample store. Enter your store URL to see both in your own brand. **Returns & claims** - **Returns portal** ([How the returns portal works](https://parcellab.com/self-service-returns-portal/)) 1. Find order: Shoppers start on your domain with their order number and email. Holiday orders get a longer return period. 2. Items: Return rules run per item and per customer: switch to a flagged customer and only store credit is offered. 3. Exchange: Exchange for another size, or for a different item at another price: the shopper pays or gets the difference. 4. Done: One QR code for drop-off and a summary of what happens next. Status emails take it from there. - **Copilot AI** ([How Copilot AI works](https://parcellab.com/copilot/agentic-returns/)) 1. Can I return it?: Returns go through the same flow as your returns portal, so the same return periods and final-sale rules apply. 2. Start a return: The return is registered in the conversation, with no form to fill in. 3. Refund status: Refund questions get the return's own status instead of a generic timeline. [Book a demo](https://parcellab.com/book-a-demo-platform-tour/) ## How the agent handles a return Four steps on every request. Each one follows rules your team sets and can change. 01 ### The request comes in A customer asks in chat, email or WhatsApp, or starts in your returns portal. The agent finds the order from live shop data. 02 ### Policy and risk are checked Each item is checked against your return windows, market rules and final-sale list, and the return gets a fraud and abuse score. 03 ### An exchange is offered first A size or color swap from live stock, or store credit with a bonus, comes before a refund wherever your policy allows it. 04 ### The return is booked The agent issues a label or QR code and sends status updates. Trusted customers can be refunded on the first carrier scan. ## One agent, the same rules on every channel Customers start a return where they already talk to you. Your policy, brand voice and returns data stay the same everywhere. ### Chat Returns and exchanges completed inside the conversation. ### Email Return and refund questions answered in your brand voice. ### WhatsApp Returns started and followed in a WhatsApp conversation. ### Returns portal Your branded self-service portal, running the same eligibility rules. ### Your own AI agent Salesforce Agentforce, Zendesk AI or Gorgias call parcelLab’s returns and claims workflows through MCP. ### Your service team Damaged items, complaints and unclear cases arrive with the order and conversation attached. ## Change the rules by asking Copilot parcelLab Copilot works inside your account. Describe a change to your return rules, check what it will touch, then approve it. [Explore Copilot](https://parcellab.com/copilot/) ### Change the setup - Extend the return window to 31 January for orders placed in November and December. - Set different return windows by market. - Restrict return carriers by destination country. - Stop a non-returnable product from getting a return label. - Add a return reason that needs a photo. ### Ask your returns data - Which products are returned most, and why? - What are our average return transit and handover times? - Compare returns portal registrations across our accounts. - Why did rejected returns spike in the first week of January? ## The agent works inside limits you set Your policy decides what the agent may approve. Anything outside it goes to a person. SOC 2 Type II certified EU AI Act compliant GDPR compliant ### Acts only when the answer is clear The agent resolves a return only when your policy and the order data give a clear answer. ### Hands over with context Damaged items, complaints, unclear cases and missing data go to your team with the conversation attached. ### Starts small Run it on a share of traffic, one brand, one market or a few case types before you widen it. ### Keeps every decision on record Transcripts and logs are there for QA and audit. Customer data never trains third-party models. ## Returns already run on parcelLab Retailers use the same portal, rules and return data the agent works from. “Returns will continue to be received in the future. With the online returns portal we now have a digital and uncomplicated solution with which we know as soon as the parcel is at the counter or with the delivery agent what is coming back and when. We benefit from this, especially in terms of staff planning and we are also faster thanks to automated handling processes.” Silvio Heider Business Expert Project & Digital Process Management, Conrad [Read the Conrad story](https://parcellab.com/case-studies/conrad-self-service-returns/) 6000 + hours saved at Conrad with self-service returns 20 % fewer WISMO inquiries at Wyze with the returns portal and post-purchase emails in one place ## Your questions, answered What retailers ask before they put an agent on returns. What is agentic returns? An AI agent runs the return for the customer: it reads the order, applies your return policy, offers an exchange or store credit, books the label and keeps the customer updated. The parcelLab Returns Agent does this in chat, email, WhatsApp and your returns portal. Does the agent replace our returns portal? No. The portal and the agent use the same rules, return reasons and return methods. Customers can start in either, and your returns data stays in one place. Can it work with the chatbot we already have? Yes. parcelLab supports MCP, so an existing agent such as Salesforce Agentforce, Zendesk AI or Gorgias can call parcelLab’s returns and claims workflows. You keep your front end. How does it choose between exchange, store credit and refund? From your rules. You set which options apply by return reason, product, market and customer, and whether store credit carries a bonus. Exchanges and store credit come first wherever your policy allows them. What stops return fraud and abuse? Every return is scored on behavior, history and risk signals before a refund is approved. Trusted customers are refunded on the first carrier scan, flagged returns wait for warehouse inspection, and your team can override any flag. When does a person take over? When a case is outside the agent’s scope: damaged items, complaints, unclear evidence or missing data. Escalation rules are set per account, and you can keep whole case types with your team during launch. How is customer data handled? The standard parcelLab MSA includes a Data Processing Agreement that covers AI sub-processors. Your data is kept separate from other customers and is never used to train third-party models. parcelLab is SOC 2 Type II certified. ## More of the post-purchase platform [Agentic order tracking Delay messages first, then answers to every order question. ](https://parcellab.com/copilot/agentic-order-tracking/) [Self-service returns portal Your branded portal, running the same return rules. ](https://parcellab.com/self-service-returns-portal/) [Claims management Missing and damaged parcels turned into carrier-ready claims. ](https://parcellab.com/claims-management/) [Fraud and abuse protection Score every return and route it by risk. ](https://parcellab.com/fraud-and-abuse/) ## See the Returns Agent on your return policy Talk to our team about your policy, your channels and the returns you want off your team’s desk. [Book a demo](https://parcellab.com/book-a-demo-platform-tour/) [Take a tour](https://parcellab.com/take-a-tour/)