AWS Advances AI Fraud Detection
Start here
AI-assisted source synthesis · not an independent investigation.
On October 27, 2026, Paul Chang, head of AWS’s payments network, said that banks are adopting machine learning, transaction foundation models, and graph analysis to identify constantly changing patterns of fraud. Transaction foundation models can generate representations from sequences of financial activity, while graph analysis is used to link accounts, transactions, IP addresses, and devices to help uncover connections related to fraud. AWS is also exploring the use of AI agents to organize investigation evidence and make recommendations. However, case disposition and the submission of suspicious activity reports (SARs) are still decided by humans.
AI-written from reports · updated 5 hours ago
Timeline
Follow the reporting to see the different sides of the story.
- PYMNTSAWS explores using AI to track fraud patterns and assist banks with investigations
Paul Chang, head of payment networks at AWS, said banks are using machine learning, transaction foundation models, and graph analytics to identify evolving fraud patterns. Transaction foundation models generate representations from sequences of financial activity, while graph analytics connect accounts, transactions, IP addresses, and devices; AI agents can organize investigative evidence and make recommendations, but case disposition and SAR filing remain human decisions.
Heat over time
Current heat 8·Peak over the comparable range 9 (Oct 27 19:00)·Change over the comparable range in the last 24 hours –
Trends compare only the same entities that were fully and continuously observed, so the range can be narrower than the current heat statistics. Move the pointer or tap the chart to read hourly heat; arrow keys work on a keyboard.