UCP Return Extension: Completing the Commerce Loop

A new extension enables AI agents to understand return policies, methods, and fees — answering "Can I return this?" before purchase.

About the Author

venkatesh-ucp contributes to the Universal Commerce Protocol specification, focusing on extending UCP's coverage of real-world commerce primitives.

The Problem

The Universal Commerce Protocol has steadily built out the forward commerce loop: product discovery via Catalog Search, cart management, discount application, and checkout. But physical commerce isn't complete at checkout — returns are a core primitive that merchants, platforms, and buyers all need to understand.

Today, when an AI agent assists with a purchase, it can't answer basic questions like:

"Can I return this if it doesn't fit?"
"How many days do I have to return this?"
"Is there a restocking fee?"
"Can I return this in-store instead of mailing it back?"

The buyer (or agent) would need to leave the commerce flow to hunt for policy information on the merchant's website — breaking the conversational experience and creating friction.

What Changed

PR #257 introduces a Return Extension to the UCP Checkout capability. The extension enables businesses to communicate return conditions, methods, timelines, and costs directly through the protocol.

Key Design Decision: The proposal explicitly argues for a standalone extension rather than embedding returns within the existing Fulfillment Extension. The reasoning: fulfillment options are selectable (the buyer chooses Standard vs. Express), while return policies are immutable rules assigned by the merchant. Mixing interactive choices with static policies creates semantic confusion.

Technical Details

The extension introduces three new schema objects:

Return Policy Response — The top-level object that maps return rules to specific line items:

return_window_type: lifetime, no_returns, final_sale, or finite_window
return_days: Number of days allowed (for finite windows)
exchanges_allowed: Whether exchanges are permitted even when refunds aren't
line_item_ids: Which cart items this policy applies to

Return Method — How the buyer can physically return items:

in_store: Drop off at retail location
by_mail: Ship back via carrier
kiosk: Self-service drop-off point

Return Fee — The cost structure for each method:

free: Merchant covers all costs
fixed_fee: Flat restocking or shipping fee
customer_responsibility: Buyer pays their own return shipping

Example payload showing mixed policies within a single cart:

{
  "return_policies": [
    {
      "id": "rp_apparel",
      "line_item_ids": ["shirt", "pants"],
      "return_window_type": "finite_window",
      "return_days": 30,
      "exchanges_allowed": true,
      "methods": [
        {
          "type": "in_store",
          "fee": { "type": "free" }
        },
        {
          "type": "by_mail",
          "fee": { "type": "fixed_fee", "amount": 500 }
        }
      ]
    },
    {
      "id": "rp_final_sale",
      "line_item_ids": ["custom_engraved_watch"],
      "return_window_type": "final_sale",
      "exchanges_allowed": false
    }
  ]
}

Why This Matters

Pre-purchase confidence. AI agents can now proactively answer return questions before the buyer commits. This reduces cart abandonment and post-purchase disputes.

True cost transparency. An agent calculating the "real" cost of a purchase can factor in return risk. A $50 item with a $15 restocking fee has different economics than one with free returns.

Omnichannel routing. The methods array enables agents to suggest the most convenient return path — "You can return this at the downtown store instead of mailing it."

Exception handling. Mixed carts with different policies (regular items + final sale items) can be surfaced clearly. The agent can warn: "Note: the engraved watch is final sale and cannot be returned."

Next Steps

The PR is under Technical Committee review. Key questions being discussed:

Whether return_days should be calculated from delivery date or purchase date
How to handle conditional policies (e.g., returns only accepted if item is unworn with tags)
Integration with the proposed eligibility claims system for premium return privileges

Combined with Catalog Search, Cart Discounts, and the Eligibility system, UCP is steadily building toward a complete agentic commerce platform where AI agents can handle the full shopping lifecycle — discovery through returns.