Monica Eaton, Founder and CEO of Chargebacks911, recently contributed a guest commentary to The AI Journal examining a critical gap in the payments industry's approach to artificial intelligence. In the piece, Monica argues that businesses have focused heavily on using AI to prevent fraud before a transaction is approved, while overlooking the equally important challenge of defending legitimate transactions when disputes arise later.
The piece explores how commerce has moved beyond the traditional checkout experience. Subscriptions, automated repeat purchases, and increasingly autonomous AI shopping agents can initiate transactions with little or no direct human involvement at the point of purchase. Monica argues that this shift requires merchants to rethink how they use AI, particularly when it comes to collecting and interpreting evidence after a transaction has been completed.
“We have spent years teaching AI how to approve payments but I argue that we’ve barely started teaching it how to defend them,” Monica writes.
The article points to findings from our 2026 Chargeback Field Report, which surveyed more than 250 merchants. While more than a quarter of respondents already use AI-based fraud prevention tools and another 37% plan to adopt them, the average net revenue recovery rate from disputed transactions was just 10.7%. Monica argues that this gap exists because traditional pre-transaction fraud tools are not designed to answer the questions that matter after a chargeback occurs, including whether a purchase was genuinely intended, whether the goods or services were delivered as promised, and whether the merchant can provide sufficient evidence to demonstrate what happened.
Monica advocates for what she calls “post-transaction intelligence,” an approach that connects evidence from payment gateways, customer service records, delivery confirmations, communications, and other sources across the transaction lifecycle. As AI becomes increasingly embedded in commerce, she argues that merchants should focus not only on deciding who should be allowed to buy, but also on whether their systems can explain and defend those decisions when a transaction is later challenged.
