Decades of mission-critical innovation inside the Department of the Air Force (DAF) have produced something every IT leader recognizes: deeply entangled, sparsely documented legacy systems that no one fully understands anymore. These systems often underpin Business and Enterprise Systems (BES) Directorate capabilities and Space Force satellite ground control and C2 software. They are too important to ignore, too fragile to change with confidence, and too opaque to migrate cleanly to cloud‑native or zero‑trust‑aligned architectures.

This session introduces a pragmatic approach to this challenge: using coordinated “teams” of Agentic AI to systematically deconstruct, analyze, and document legacy code bases at scale—turning opaque mission software into a well‑understood, refactorable asset. Unlike single large language models that passively answer questions, Agentic AI uses a team of specialized, collaborating agents with defined roles, goals, and workflows. In the legacy modernization context, these agents crawl and inventory mixed-language repositories, infer architecture and data flows, cluster related components into mission-aligned functional domains, generate living documentation (sequence diagrams, data dictionaries, interface definitions, dependency graphs), and surface technical debt hot spots such as duplicate logic, dead code, brittle integrations, and security anti‑patterns.

The session will walk through an end‑to‑end modernization scenario: from discovery and ingestion of millions of lines of code and artifacts, through semantic reverse engineering and architecture reconstruction, to human‑in‑the‑loop governance and feeding the cloud‑native pipeline (backlogs, CI/CD, and model‑based engineering). We will spotlight early lessons learned—where Agentic AI excels in pattern detection and breadth‑first analysis, and where human expertise is indispensable for mission intent, risk tradeoffs, and operational edge cases.

The focus is not on AI replacing engineers, but on eliminating the archaeological drudgery that precedes every modernization effort. Attendees will leave with a concrete playbook for reducing technical debt, de‑risking modernization of mission‑critical legacy systems, and establishing a repeatable Agentic AI framework to support BES and Space Force refactoring initiatives—turning the “ghost in the code” into a clear blueprint for next‑generation DAF cloud‑native, AI‑enabled capabilities.

Key Learning Objectives

By the end of this session, participants will be able to:

• Describe how Agentic AI teams can systematically analyze legacy DAF systems by inventorying mixed-language code bases, inferring architectures and data flows, and generating living documentation.

• Explain the end‑to‑end AI‑assisted modernization workflow from discovery and ingestion, through semantic reverse engineering and architecture reconstruction, to integration with agile backlogs and CI/CD pipelines.

• Identify where Agentic AI delivers the most value (e.g., pattern detection, dependency mapping, technical debt discovery) and where human engineers must remain central (mission intent, risk decisions, edge cases).

• Assess how AI-generated artifacts support cloud‑native and zero‑trust transformations by providing target‑agnostic architecture views, interface catalogs, and domain models suitable for refactoring to containers, microservices, or serverless patterns.

• Apply a practical playbook to reduce technical debt and de‑risk refactoring efforts, including governance practices that ensure traceability, cybersecurity compliance, and repeatable use of Agentic AI on BES and Space Force legacy portfolios.