For teams maintaining large OpenEdge ABL systems.
Agentic Doesn't Mean Autonomous
Asking an LLM to rewrite legacy ABL leaves you with the same coupling, no evidence that behavior is unchanged, and a diff too large to review. This session shows a controlled agentic workflow built on a synthetic OpenEdge 12.8.12 application, including the mistakes the AI made and what they cost in tokens and time.
What you'll see
1. Understand
Inventory, XREF dependency graph, behavior, and per-file risk for the application.
2. Decide
Pick the first seam using cited guidance from Progress Agentic RAG, with human approval tied to an exact plan.
3. Change
Modify only the files in the approved plan.
4. Prove
Compilation, legacy characterization tests, PASOE deployment, and an HTTP contract check.
Talk through your OpenEdge modernization
Bring one application. We'll look at where a controlled, evidence-backed approach fits and where it doesn't.
Book a modernization conversation