SQL Server to Aurora PostgreSQL conversion with AI agents for AWS DMS

This article demonstrates how to use AI agents with AWS DMS Schema Conversion to automate SQL Server to Aurora PostgreSQL migrations using natural language orchestration.

• DMS Schema Conversion processes objects through a four-step pipeline: parse, resolve, transform, and generate (with optional AI assistance)
• Deterministic rule-engine conversions produce consistent, repeatable output; generative AI-assisted conversions require human review and functional testing
• Action item severity levels (INFO, LOW, MEDIUM, HIGH, CRITICAL) guide remediation effort and risk assessment
• Agents orchestrate multi-step workflows including metadata import, conversion, export, and report generation with natural language prompts
• Decision frameworks address common incompatibilities: full-text search, error handling, CLR assemblies, and hierarchyid functions
• Scale operations safely by serializing dependent tasks, respecting AWS DMS quotas, and using bounded concurrency across independent projects

Successful migration requires validating converted code through regression testing and functional verification before production deployment, regardless of conversion method or action item severity.