Details
- Thought Machine and Amazon Web Services announced an AI-powered solution on 29 September 2026 to accelerate banks’ migration from legacy mainframe systems.
- The solution combines AWS Transform’s mainframe reverse-engineering agents with Thought Machine’s Vault Forge, powered by Amazon Bedrock.
- Its pipeline extracts business rules from COBOL and other legacy code, consolidates duplicate product variants, synthesises SDK-compliant Python financial products and test suites, and validates results through human review in a secure AWS environment.
- Unlike line-by-line code translation, the approach converts business intent into cloud-native product logic designed for Thought Machine’s Vault platform, where financial products are decoupled from core infrastructure.
- Thought Machine is also launching a Core Modernisation Accelerator programme, with architects, product specialists and engineers supporting banks through governance, architecture and implementation.
Impact
The announcement reinforces a broader shift in mainframe modernisation toward agentic AI, business-rule extraction and cloud-native reimplementation rather than direct code conversion. If validation and governance controls perform as described, banks could shorten discovery and prototyping cycles, while demand grows for AI-assisted migration expertise and platforms that preserve auditability in regulated environments.
