AWS re:Invent 2025 - Accelerate Legacy Modernization with Slalom & AWS (MAM216)

Accelerating Legacy Modernization with Slalom & AWS

Defining Legacy Systems

• Legacy systems are defined as any technology or system that poses a barrier to modernization and innovation
• Key characteristics of legacy systems:
  • Monolithic and stateful architecture
  • Mainframe or AS/400 workloads (e.g. z/OS, COBOL)
  • Difficulty hiring skilled talent to maintain the systems

Drivers for Legacy Modernization

• Rising costs and challenges hiring skilled talent to maintain legacy systems
• Opportunity to leverage AI and generative AI technologies to reverse engineer and understand legacy codebases
• Ability to use generative AI to boost developer productivity and enable spec-driven development

Slalom's Approach to Legacy Modernization

1. Reverse Engineering with Generative AI:

   • Use tools like Amazon Kendra and Bedrock to quickly index and understand legacy codebases
   • Leverage generative AI models to extract high-level domain knowledge and execution paths
   • This enables a more efficient re-engineering process compared to manual documentation

2. Augmenting the SDLC with Generative AI:

   • Use AI-powered coding assistants like Amazon Q to boost developer productivity by 30% or more
   • Integrate the reverse-engineered legacy context into the AI-assisted development workflow

3. Spec-Driven Development:

   • Leverage the detailed specifications extracted from the legacy system
   • Use tools like Kira CLI to enable a spec-driven development approach
   • Ensures full traceability between the new system and the legacy codebase

4. Stakeholder Alignment and Organizational Change:

   • Engage all key stakeholders across the organization
   • Align on the modernization strategy and key outcomes
   • Ensure a holistic approach that addresses both technical and organizational challenges

Lazy Boy Case Study

• Lazy Boy had a mainframe-based legacy system that was difficult to maintain
• Slalom conducted an AWS-funded assessment to fully understand the legacy system
  • Mapped dependencies, engaged stakeholders, and identified a path forward
• Used a phased approach to modernize the system:
  • Leveraged Amazon Q and step functions to accelerate infrastructure deployment
  • Transformed COBOL to Python and migrated the DB2 database to Amazon Aurora PostgreSQL
  • Implemented a rigorous testing process to ensure functional equivalence
• Uncovered previously unknown bugs that led to business process improvements
• The modernization enabled Lazy Boy to break free from the constraints of their legacy system and explore new opportunities

Key Takeaways

• Generative AI is a game-changer for reverse engineering and understanding legacy codebases
• Integrating AI-powered tools into the SDLC can drive significant productivity gains
• A holistic, stakeholder-aligned approach is critical for successful legacy modernization
• Modernization can uncover hidden issues and enable broader business process improvements
• Slalom's "zero legacy" approach helps clients break free from the constraints of legacy systems and unlock new opportunities