Across enterprise commerce, fraud no longer behaves like a series of isolated incidents. It operates as repeatable, automated playbooks — continuously tested, refined, and redeployed across environments. The result is not just more fraud, but fraud engineered to blend in with legitimate customer behavior. That shift is changing where fraud risk sits, how it needs to be managed, and what it costs when it isn’t.

This report outlines five themes defining fraud strategy in 2026:

1. Fraud has become systematic. Automation has turned fraud into a continuous test-and-learn cycle. Tactics are deployed, refined in real time, and scaled quickly, with successful approaches repeated across environments and targets.

2. Good customers have learned to game the system. First-party fraud and policy abuse increasingly operate within legitimate customer journeys. Merchants no longer need to only verify identity; they also need to understand intent.

3. Precision drives growth. The cost of stopping fraud is increasingly measured in lost customer lifetime value, not just prevented loss. False declines, rising manual review costs, and blanket controls suppress growth. 

4. Identity cannot be a static indicator. Point-in-time checks are no longer sufficient. Trust is built over time through behavior, history, and recognition across sessions, devices, and environments.

5. Trust must extend beyond the transaction. As AI agents begin to act on behalf of customers, fraud prevention must start earlier, with the systems that govern behavior before payment occurs.

Chapters 1–3 establish the nature and scale of these shifts. Chapters 4–5 address how leading organizations are responding — and where the greatest leverage now exists.

The organizations pulling ahead aren’t those with the strictest controls. They’re those with the most precise ones.

Methodology

For this report, we retrieved transaction data from the Adyen platform for the full year 2025 (US$1.6T in data) and separately surveyed 1,000 US enterprise merchant decision makers. Survey data in the report is labeled as Adyen survey data. Platform data is labeled as Adyen platform data .