FORMAT: 45-Minute Presentation + Q&A | AUDIENCE: Program Managers, Enterprise Architects, CIOs, DevSecOps Practitioners | CLEARANCE: Unclassified / FOUO

FULL SESSION DESCRIPTION

The Air Force Business Enterprise Systems (BES) portfolio carries decades of accumulated technical debt of aging .NET, Java, and Mainframe applications with undocumented architectures, orphaned interfaces, and database schemas that exist only in the institutional memory of engineers long since departed. Eighty percent of DoD IT budgets are consumed sustaining these legacy systems, more than sixty percent of legacy codebases are undocumented or orphaned and change failure rates run three to five times higher than modern applications. Programs that attempt modernization manually face six to twelve months of code archaeology before a single line of modern code is written which slows down modernization. This session presents a practical, responsible AI-driven framework that fundamentally changes the economics, speed, and risk profile of legacy system modernization across the BES portfolio.

The Problem

Legacy BES applications share a common and well-recognized set of failure patterns. Monolithic .NET, Java, and COBOL/Mainframe architectures are tightly coupled, resistant to change, and structured in ways that make incremental modernization extraordinarily difficult. Requirements documentation has not kept pace with years of patches, workarounds, and emergency fixes. Interface specifications for downstream consumers, REST and SOAP integrations, message queue patterns, and file-based data exchanges frequently exist nowhere in writing, surfacing only as production incidents during migration sprints. Aging relational databases like DB2, Oracle, Sybase carry embedded business logic in stored procedures that no current team member fully understands. Test coverage is too thin to safely support migration, meaning every change carries unquantified risk. The cumulative effect is a modernization environment where the cost and risk of change is so high that programs defer action until the system becomes a mission liability.

The AI-Driven Approach

This session demonstrates how AI can be applied across the full modernization lifecycle as a force multiplier that eliminates the most time-consuming and error-prone phases of the process.

Automated Code Discovery and Dependency Mapping: AI ingests .NET, Java, and COBOL/Mainframe source code and constructs complete Abstract Syntax Tree-based call graphs, module inventories, and cross-system dependency matrices. Every class, function, stored procedure, and shared component is catalogued with AI-generated plain-language summaries. The result is a risk-ranked dependency map that makes the blast radius of any proposed change visible before a single line is touched compressing months of manual code archaeology into hours.

Undocumented Interface and Integration Detection: AI identifies REST and SOAP calls, MQ message exchange patterns, file-based integrations, and database-level coupling that exist nowhere in documentation. These hidden dependencies are the leading cause of unplanned outages during modernization sprints. Surfacing them before migration begins is the single highest-leverage risk reduction available to a program.

Requirements Reconstruction from Legacy Code: Large Language Models extract implicit business rules from variable names, code comments, conditional logic, and execution patterns, regenerating functional specifications from the codebase itself when original requirements document either doesn’t exist or reflect reality.

Auto-Generated Architecture Diagrams: C4 context, container, and component diagrams are produced directly from source code analysis, validating or replacing missing architecture artifacts without the months of reverse-engineering workshops that traditional approaches require.

AI-Assisted Refactoring - Monolith to Microservices: Domain clustering algorithms and LLM analysis identify cohesive bounded contexts within monolithic codebases, proposing microservice decomposition boundaries aligned to business capability rather than technical convenience. The Strangler Fig pattern is supported with AI-recommended routing and proxy layer strategy, enabling parallel operation of legacy and modern components during incremental cut-over.

Technical Debt Identification and Scoring: AI code analysis engines score security vulnerabilities by CVSS severity, cyclomatic complexity, deprecated API usage, duplicate code blocks, and missing error handling, producing a PM-consumable risk-ranked remediation roadmap that allows program leadership to make sequencing decisions based on mission impact rather than technical preference.

AI Test Case Generation and Coverage Acceleration: AI synthesizes unit, integration, and regression test suites from existing code paths and production execution traces will dramatically increase test coverage before and after migration, and enabling CI/CD pipeline gates that enforce coverage targets automatically.

Containerization and Cloud Migration Acceleration: AI analyzes runtime dependencies and generates production-ready Dockerfiles, Kubernetes manifests, Helm charts, and Terraform modules eliminates weeks of hand-crafted infrastructure configuration. AI-generated CI/CD pipelines for GitHub Actions and Azure DevOps include integrated security scanning, SAST/DAST gates, and automated rollback logic. DoD cloud compliance gap analysis maps application behaviors against CC SRG and NIST 800-53 control families, flagging gaps and generating POA&M inputs before ATO submission. These capabilities compress cloud onboarding timelines in IL2 through IL5 environments by an estimated forty to sixty percent.

The Role of Mission Experts

AI can identify code patterns, surface dependencies, propose refactoring options, and generate test cases but mission-experienced technologists remain essential at every stage. Air Force subject matter experts validate operational impact, cybersecurity risk, user workflows, data sensitivity, and production readiness. No AI output in this framework is self-executing. Every AI recommendation is a proposal that a human architect, developer, or program manager evaluates, challenges, modifies, or approves. The framework is designed to make expert judgment faster and better-informed.

Responsible AI Governance

All AI capabilities presented operate under a responsible AI governance model aligned to the five DoD AI Ethics Principles: Responsible, Equitable, Traceable, Reliable, and Governable. Human authority is retained at every decision point, no automated code change is promoted without structured human review gates. Every AI-generated artifact carries a plain-language rationale and a complete audit trail from AI suggestion to implementation decision, supporting compliance review, ATO evidence packages, and program audits. AI tooling operates within approved government cloud enclaves; no proprietary source code is transmitted to commercial LLM APIs without data protection controls in place. Modernization priorities are filtered through mission impact scoring defined with Air Force stakeholders.

SESSION STRUCTURE (45 Minutes)

• 0–5 minutes: Scene setting, the Air Force legacy modernization status quo, why traditional approaches fail, and what responsible AI changes about the calculus
• 5–15 minutes: Live walkthrough of AI code discovery, dependency mapping, interface detection, and auto-generated architecture diagram production from actual legacy source
• 15–30 minutes: AI-assisted monolith decomposition, technical debt scoring and roadmap generation, and AI test suite synthesis, with human review gates demonstrated at each step
• 30–40 minutes: Containerization acceleration, cloud migration pipeline generation, DoD compliance gap analysis, and responsible AI governance model in practice
• 40–45 minutes: Q&A and audience discussion

LEARNING OBJECTIVES

Upon completing this session, attendees will be able to:

1. Identify AI-appropriate modernization touchpoints within legacy .NET, Java, and Mainframe BES application portfolios and articulate the business case to program leadership (Primary audience: Program Managers, Program Executives)
2. Apply AI-driven code discovery and dependency mapping to produce complete architecture diagrams and interface inventories without relying on existing documentation (Primary audience: Enterprise Architects, Technical Leads)
3. Use AI-assisted refactoring and test generation techniques to reduce migration risk when decomposing monolithic systems into microservices (Primary audience: DevSecOps Engineers, Developers)
4. Evaluate and score technical debt using AI code analysis tools and translate findings into a risk-ranked, PM-consumable modernization roadmap (Primary audience: Program Managers, Architects, CIOs)
5. Apply a responsible AI governance model ensuring human oversight, full audit traceability, and alignment with DoD AI Ethics Principles throughout the modernization lifecycle (Primary audience: All roles)

KEY TOPICS COVERED

Automated code discovery and dependency mapping · AI-assisted monolith-to-microservices refactoring · Technical debt identification and risk scoring · Undocumented interface and integration detection · AI-generated architecture diagrams (C4 / UML) · Requirements reconstruction from legacy code · AI test case generation and coverage acceleration · Containerization and cloud migration acceleration (IL2–IL5) · Responsible AI governance and human oversight · DoD AI Ethics alignment, TEVV practices, and ATO evidence support

DAFITC THEME ALIGNMENT

Accelerating Mission Readiness Through Technology -> AI compresses legacy analysis from months to hours, directly accelerating program schedules that drive readiness

Modernizing the Air Force Enterprise -> A replicable framework applicable across the entire BES portfolio

Responsible AI Integration in Government IT -> Demonstrates how DoD AI Ethics Principles translate from policy to engineering practice in a real modernization context

DevSecOps and Cloud Adoption at Scale -> AI-generated IaC, Dockerfiles, and CI/CD pipelines operationalize DevSecOps for programs that have never had automated delivery

Workforce Upskilling for AI-Augmented Development -> Attendees leave with practical AI tool literacy applicable immediately to their programs

ORGANIZATIONAL CREDENTIALS

ITC Federal brings direct experience modernizing Air Force and DoD enterprise systems, bridging legacy environments and cloud-native, mission-ready architectures. Relevant experience includes:

• USAF DCGS (Prime Contractor): Delivered COTS Sustainment Support including enterprise configuration management, logistics, and technical sustainment for the globally distributed ISR weapon system across 8 sites with TS/SCI-cleared personnel, providing the mission-system operational context that informs this framework
• USCIS - AI-Enabled Modernization: Deployed the ELIS Assistant AI chatbot inside a live mission environment; supported Amazon EKS, CloudBees Jenkins, Terraform, and Splunk with weekly release governance; reduced average incident resolution time by 20% and cut open incidents from more than 8,000 to fewer than 6,000
• ICE OCIO - Multi-Cloud DevSecOps: Supports a multi-cloud environment spanning 100+ applications and 90+ CI/CD pipelines, achieving 99.9% availability through SRE-based operations across 600 sites

ITC Federal maintains an established AI/ML practice with active deployment of LLM-assisted code analysis, test generation, and IaC scaffolding in federal environments. Organizational policies and delivery frameworks are aligned to DoD AI Ethics Principles, with all AI tools operating under human oversight and full audit traceability.

ATTENDEE TAKEAWAY

Attendees leave with a replicable, responsible AI modernization playbook, concrete techniques, governance guardrails, and a sequenced approach applicable to any BES portfolio application, grounded in ITC Federal's direct experience delivering AI-enabled modernization across federal mission environments.