Agentic AI · Cloud Modernization
Accelerating SQL Server database modernization
An AI-assisted workflow for full-stack SQL Server to Aurora PostgreSQL transformation inside AWS Transform, launched at re:Invent 2025. I defined the conversational framework and the end-to-end modernization journey.
Final result
A guided workflow that lets cross-functional teams automate transformation end to end — layered from chat summary, to structured data table, to drill-down detail.
Companies still run applications coded twenty years ago, paying millions in licensing. When they commit to modernizing, 80% feel uncertain: assessments are inaccurate, planning is guesswork, and 90% of schema conversion still needs manual work even with automation.
How might we guide users through a SQL modernization journey with clarity, confidence and control — while interacting with multiple AI agents?
Rebuilding the conversational frame
The existing framework was dense and flat; users could not tell where to click to interact with the agent. I worked with the platform team on a three-panel model: the job plan on the left, chat as the primary surface, and a human-in-the-loop panel that collects the decisions actually needing a person.
Agent as partner, trust but verify
Every agent output is reviewable before it executes. Chat handles intent, explanation and quick edits; the UI holds structured data, dependencies, bulk operations and validation. Configuration happens step by step so teams can work waves concurrently without inheriting each other's uncertainty.
Outcome
Featured as a key announcement at AWS re:Invent 2025 as evidence of agentic AI accelerating enterprise cloud migration.