The Ship-Date and Refund Promises Your Store’s Website Shouldn’t Make

A shopper with a cart full of gifts asks your store’s website assistant two questions: will it arrive by Friday, and are returns free? The bot says yes to both. The warehouse needs five business days, and your policy charges return shipping on opened items. The order ships late, the refund comes back short, and the customer files a chargeback with a screenshot of the chat attached. The store wanted a tool to reduce support tickets. It just manufactured one, and bound itself to promises it cannot keep.

Online retail feels like the safest place to drop in a chatbot. It is not. Shipping timelines fall under one federal rule. Refunds answer to another. Subscriptions and product claims each carry their own. A general-purpose bot fields questions across all of it and knows none of the law sitting behind any of it. Every confident answer becomes a representation the store has to stand behind.

A Ship Date Is a Federal Promise

When a shopper asks "will it get here by Friday," the honest answer depends on warehouse stock and carrier timing the website never actually sees. A generic bot answers anyway. The trouble is that the FTC’s Mail, Internet, or Telephone Order Merchandise Rule sets real obligations around shipment timing and what a seller must do when it cannot ship when promised. A chatbot that casually commits to a delivery date the operation cannot meet does more than disappoint a customer. It can put the store crosswise with the rule that governs exactly that promise.

Refunds are the same story. "Returns are free" and "you’ll get a full refund" are policy statements, and when the bot’s version contradicts the actual policy, the store is caught between honoring a promise it never authorized and defending one its own site put in writing. The FTC’s authority over deceptive acts lives precisely in that gap.

Subscriptions and Product Claims Carry Their Own Rules

If you sell anything on a recurring plan, a bot that explains the subscription loosely runs into the negative-option requirements of the Restore Online Shoppers’ Confidence Act, which dictates how recurring charges must be disclosed and cancelled. Product claims are the other trap. A bot that promises what an item does, or how fast it works, or how much a shopper will save, can hand a regulator a deception claim. The reason is simple: a substantiation standard applies to that promise no matter who on the site made it. The store did not approve a marketing message. A widget improvised one.

The Email Trap

A bot that offers to "sign you up for deals" or fires off follow-ups can drag the store into CAN-SPAM’s rules for commercial email, from accurate headers to a working opt-out. Consent and compliance there are the store’s responsibility, even when an automated assistant set the messages in motion.

The Price the Bot Quotes Can Become the Price You Owe

Pricing is where it gets expensive. A shopper asks whether a discount still applies. The bot does the math and gets it wrong, or honors a promotion that expired last week. Now a confirmed price is sitting in a transcript, and the customer relied on it. Clawing it back after checkout reads as exactly the deceptive practice the FTC polices.

At volume, an honest glitch stops being a glitch. The bot makes the identical mistake for everyone who asks, so one bad pricing answer is a thousand refund requests by morning. Then there is the data question. These conversations also gather personal information: an email here, a shipping address there, a running record of past orders. A bolt-on widget pipes all of it wherever it happened to be wired. A store that owes its shoppers any duty of care over that information now answers for where it ended up. Fewer tickets was the pitch. A faster way to be wrong in writing was the result.

Who Owns the Answer

Strip away the software and the exposure is ordinary retail law. A support rep who promises a delivery date the warehouse cannot hit, or a refund the policy does not allow, creates a problem the store has to clean up, and good operations train against it. An unsupervised chatbot makes those promises at scale, in writing, with no one reviewing them. The accountability does not move to the vendor. It stays with the store, now with a transcript a customer can wave at a chargeback team or a regulator.

The merchants who get burned are the ones who assumed a helpful bot was a harmless one. The fix is not to pull AI off the storefront. It is to run one that treats a ship date, a refund, a subscription term, and a product claim as the regulated promises they are. That assistant sends each to policy and to people, rather than inventing an answer a shopper will later hold you to.

Frequently asked questions