# Merchants: how do you plan to handle disputes on purchases made by AI agents? — retained record

> Retained reddit record from reddit:fintech.

Canonical: https://highsignal.app/data/records/8b06ff38bc05a21f

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:fintech)  published 2026-11-02 15:04 ingested 2026-11-03 02:43

[original source ↗](https://www.reddit.com/r/fintech/comments/1xlzkp2/merchants_how_do_you_plan_to_handle_disputes_on/)

## retained text



```
Agent-driven checkout is starting to show up (AP2, Visa and Mastercard agent programs, Stripe, etc.), and a lot of what I read says liability for these purchases is still unsettled, so merchants may end up absorbing the loss unless they can show the customer authorized the agent and it stayed in scope.



I'm curious how this looks from the merchant / payments ops side, not the theory:



1. Have you already had a dispute or a customer saying "I didn't buy that, my AI did"? What happened?

2. What evidence would you want on hand for an agent-initiated order (who authorized the agent, what limits it had, what it actually did)?

3. Would you trust a record kept by the agent's own platform, or would you want something verifiable by a neutral party?
```



document fetched 2026-11-03 00:46 · canonical https://reddit.com/r/fintech/comments/1xlzkp2/merchants_how_do_you_plan_to_handle_disputes_on
