AWS re:Invent 2025 - Accelerate Telco Transformation: AT&T's AI-Powered Migration at Scale (IND201)

Accelerating Telecom Transformation with Generative AI

Challenges of Telecom Modernization

• 70% of telecom IT still runs on-premises, with 70% of that being over 20 years old
• Migrating these critical systems (e.g. billing, customer management) takes 1.5+ years on average
• Key challenges include:
  • Mission-critical availability requirements (e.g. 99.999% uptime)
  • Massive data volumes and complex business logic
  • Strict regulatory compliance (e.g. CPNI, PCIDSS, GDPR)
  • Highly integrated legacy systems and technical debt
• Traditional migration approaches are time-consuming, require reverse-engineering, and are error-prone

Generative AI for Transformation

• Generative AI has evolved from simple chatbots to autonomous "agents" that can reason, act, and learn
• Agentic AI uses a cycle of observing, reasoning, acting, and reflecting to execute complex workflows
• Multi-agent systems can collaborate to handle end-to-end transformation processes

AWS Agentic AI Services

• AWS Transform: Industry-purpose built service for automated migration of legacy systems (e.g. VMware, .NET, mainframe)
  • Includes code analysis, documentation generation, code decomposition, and refactoring agents
• Amazon Q Developer: Specialized agents for software development tasks like code changes, testing, documentation
  • Includes agents for code upgrades, unit testing, code review, and transformation
• AWS Keiro: Allows reimagining legacy apps as modern microservices-based applications
  • Starts from requirements, not just code, to generate new cloud-native architectures

AT&T's Mainframe Modernization Journey

• Challenges include lack of mainframe expertise, highly integrated legacy systems, and need to maintain business continuity
• Traditional rehosting or code conversion approaches were insufficient
• Leveraging AWS agentic AI services to:
  • Automatically analyze and document legacy mainframe applications
  • Decompose monoliths into microservices architecture
  • Refactor and transform code with automated testing
  • Reimagine legacy functionality as new cloud-native services
• Goal is to accelerate migration velocity, improve performance, and de-risk the business

Key Takeaways

• Generative AI is redefining legacy modernization, enabling 70%+ automation
• AWS provides a suite of agentic AI services to handle end-to-end transformation
• AT&T is leveraging these capabilities to rapidly modernize critical mainframe systems
• Reimagining legacy functionality as cloud-native microservices is a game-changer
• Automated analysis, documentation, and testing are critical to overcoming legacy challenges