Migrate and Modernize Legacy Workflows to Intelligent Agents (PEX304)
summary
Concrete roadmap for turning a SOAP legacy app into an agentic system: use Kiro code learning to extract business flow, auto-generate specs and task lists, refactor SOAP→REST and Aurora PostgreSQL; build multi-agent graphs with Strands Agents SDK and deploy on Amazon Bedrock/AgentCore using AgentCore Gateway (MCPfy) to expose tools; demo shrinks a 12-step order workflow to 8 steps via WhatsApp (AWS end‑user messaging), uses Claude Sonnet 4/Haiku/Nova Lite per agent, adds short/long‑term memories and Arize/Amazon observability, and supports 8‑hour runtimes.
AI-written summary · about
chapters (8)
- 0:00Introduction and speakers
- 3:53Why modernize workflows
- 11:55Legacy app transformation with Kiro
- 22:11Designing intelligent agents
- 37:31Multi-agent demo walkthrough
- 55:41Scaling agents for enterprises
- 1:03:00Real-world agentic success stories
- 1:06:19Key takeaways and call to action
related talks
50mHarnessing Event-Driven and Multi-Agent Architectures for Complex Workflows in Gen...
43mArchitecting the Future: Modern Solution Design with Agentic AI
51mOptimizing Complex Workflows with Event-Driven Multi-Agentic Approach
35mUnleash Your Inner API Architect: Build Agents That Deliver
25mBeyond Orchestration: An Open Source Architecture for Agentic Systems
28mBuilding Agentic AI applications on AWS
33mAgent Experience
40mHow To Tame Your Agents?