Amazon Web Services said Bank of Ayudhya PCL (Krungsri) — Thailand's fifth-largest bank, with operations across six Southeast Asian countries — is cutting cloud migration time by more than half using generative AI agents on AWS, in a Bangkok release dated 29 April 2025.
Custom GenAI agents on Amazon Bedrock, with SageMaker, Lambda and EventBridge, automatically convert on-premises programming scripts and code to cloud environments for data platform, warehouse and ML model migration — more than 50% faster versus manual conversion. Reusable AI assets including conversion algorithms, trained models and automation scripts support future migrations and ETL transforms, with ML models migrating to Amazon SageMaker. On fraud, SageMaker-based proactive mule-account detection aligns with Thai regulations banning mule accounts. Property valuation on SageMaker since February 2024 was said to have eliminated more than 2,000 on-site assessments, saved about 5 million Baht, and cut evaluation from three days to 15 minutes per case. ATM cash optimisation consolidated 6,000 ML models into three on SageMaker for cash demand across more than 5,000 ATMs.
Tul Roteseree, head of Enterprise Data and Analytics Group at Krungsri, said: "By leveraging generative AI and machine learning on AWS, we are cutting cloud migration time in half and transforming our systems into cloud-powered solutions." Performance and savings figures are company statements; AI-360 has not independently verified them.