“We’re moving from AI as a feature to AI as a workforce. These agents don’t just analyze code. They act on it.”
This issue of Redeployed is brought to you by Tecla: AWS agents can now refactor COBOL into Python while you sleep. But they don’t bring context. They don’t know which business rule saved your company in 2008.
Modernization still needs humans. Tecla helps you hire nearshore engineers fluent in AWS, legacy systems, and AI workflows, so the bots don’t rewrite your core logic into nonsense.
For years, legacy systems have been the enemy of progress. Bloated mainframes. Tangled Java logic. Thousands of lines of undocumented COBOL stitched together by engineers long since retired. Every CIO knows the drill. Everyone agrees it needs modernization. Nobody wants to touch it.
AWS just gave them a reason to.
This week, Amazon quietly expanded its Transform service to include a new category of AI functionality: agentic modernization. These are not assistants. They are code-level actors. Tasked with analyzing old systems, extracting business rules, generating updated versions in Python, Java, or Node.js, and running automated tests along the way.
In short, AWS just handed enterprise tech leaders a team of bots that can read, translate, and rebuild legacy apps. No offshore RFP. No multi-year rewrite. Just upload the codebase and let the agents start crawling.
To call this a shift would be an understatement. Because this isn’t about cloud migration. It’s about system transformation, powered by autonomous agents that work 24 hours a day and never ask for clarification.
And the targets aren’t startups. They’re banks. Insurers. Logistics companies. Government orgs. The kind of institutions where change usually means meetings, not deployments.
Of course, this isn’t new territory. Tools like Glean and Modern Systems have offered rule-extraction and code conversion for years. What’s different now is intent. AWS isn’t positioning this as a one-time migration tool. It is framing it as a living part of enterprise modernization. AI agents that integrate into your CI/CD flow. Agents that refactor. Test. Comment. Track business logic over time.
And the implications are massive.
If AI can handle legacy conversion, what happens to the contractors who have built careers around manual porting? If agents can continuously modernize a codebase, who owns the architecture decisions that follow? If business rules are extracted by a model, who verifies their accuracy?
We’re entering uncharted territory. And the senior engineers who used to hold the institutional knowledge are now being asked to review AI-generated rewrites of systems they originally hand-built.
For all the hype around automation, modernization still needs humans. Especially ones who understand AWS tooling, legacy systems, and how to work alongside AI. That’s why some enterprise teams are pairing agentic services like Transform with nearshore developers who can validate outputs, refactor intelligently, and move legacy rewrites into production. These aren’t generalists. They’re AWS-fluent engineers who operate in your time zone and already know how to co-build with automation. See how they’re accelerating modernization here.
Some will welcome the help. Others will worry the machine is rewriting the logic without understanding the business.
Both reactions are valid. Because these agents don’t bring context. They bring pattern recognition. And what looks like a legacy data transformation to an LLM might actually be a hard-coded business exception that saved a billion dollars in 2008.
Still, the direction is clear.
AWS has made it easier to act on the modernization mandate. And for CIOs with tech debt measured in millions of lines, that may be all the justification they need.
More to come…
Recommended Reads
✔️ “How We Use AI Agents for COBOL Migration and Mainframe Modernization” — Microsoft Blog
✔️ “Modernizing Legacy Code With AI” — Crowdbotics
✔️ “AWS Transform adds AI agents to modernize codes and apps faster” — TechBuzz.ai
– Gino Ferrand, Founder @ Tecla