AWS re:Invent 2025 - Accelerate .NET application modernization with generative AI (DVT211)

Accelerating .NET Application Modernization with Generative AI

Modernization Challenges and Opportunities

• Customers face significant challenges with .NET application modernization, including:
  • High Microsoft licensing costs (2x for applications, 4x for databases)
  • High costs of moving to open-source alternatives
  • Long timelines to realize ROI from modernization projects
  • Accumulating technical debt that blocks innovation and AI integration
• The modernization journey has evolved:
  • In the past, customers focused on lift-and-shift to the cloud, postponing application modernization
  • Now, generative AI tools allow application modernization to be part of the migration process
  • This new approach eliminates technical debt upfront and enables cost-saving cloud technologies

Benefits of Modernizing .NET to Linux

• Cost savings: 40% less to run Linux vs. Windows, including licensing and performance
• Performance: Linux servers and applications typically 1.5-2x faster than Windows
• Scalability: 50% more scalable on Linux
• Compatibility with ARM64 (AWS Graviton processors) and cloud-native technologies like Kubernetes and serverless

Why Are .NET Modernization Projects Stuck?

• Modernization process requires senior engineers with deep legacy .NET knowledge
• These engineers are typically assigned to feature-building projects, not modernization
• Coding assistants like Kira still require engineers to get the full benefit

AWS Transform for .NET: An Asynchronous Modernization Service

• Automates 40-70% of .NET transformation tasks, reducing the need for senior engineering resources
• Combines a rules engine with generative AI to provide a deterministic and continuously improving transformation process
• Offers both an IDE experience in Visual Studio and a web-based bulk transformation experience
• Runs the transformation in the cloud, not on the developer's machine, allowing asynchronous and scalable processing

Operationalizing .NET Modernization at Thompson Reuters

• Thompson Reuters had a mandate to use generative AI to accelerate and automate developer workflows
• Challenges included scale, fear of changing legacy code, resource contention, and building the business case
• Deployed AWS Transform for .NET and Kira to:
  • Accelerate modernization efforts, reducing months to weeks
  • Free up developer resources from maintenance to focus on new features
  • Establish a central team to operate the modernization process at scale

Key Lessons Learned

1. Start small to work through end-to-end processes and integration challenges
2. Expect more complexity with older codebases
3. Build a central team to execute modernizations and scale the process
4. People and processes are more critical to success than the technology itself
5. Develop a strategy to keep applications up-to-date and prevent future technical debt

Conclusion

• AWS Transform for .NET and generative AI tools like Kira enable organizations to accelerate .NET modernization efforts
• By automating transformation tasks and freeing up developer resources, companies can focus on building new features and driving business innovation
• Successful modernization programs require a balance of technology, people, and processes, as demonstrated by Thompson Reuters' experience