5 lessons running AI agents in production
What happens when your agentic AI prototype works great with 12 tools — but falls apart at scale? Genesys (https://www.genesys.com/) found out the hard way, then rebuilt with Strands Agents to serve 2 million users across 21 regions. In this episode, Romain sits down with John Sexton and Aaron Tummon from Genesys to discuss how they built Genesys Cloud Copilot — an agentic AI assistant that helps contact center admins and supervisors manage complexity through natural language. 🔑 Key topics covered: Why Bedrock inline agents hit a wall at 12-15 tools and how Strands Agents solved it Multi-agent architecture: orchestrator + stateless sub-agents (agents-as-tools pattern) Context management strategies: summarizing vs. sliding window conversation managers Cost optimization: moving from Sonnet to Haiku 4.5 + prompt caching (massive savings) Testing agentic systems: LLM-as-judge + integration tests + the "Sentinel" rotation Data sovereignty challenges when deploying AI across 21+ regions How domain experts contribute sub-agents without needing to understand the full framework 🔗 Resources mentioned: Strands Agents SDK: https://strandsagents.com/ Strands Agents on GitHub: https://github.com/strands-agents/sdk... Amazon Bedrock AgentCore: https://aws.amazon.com/bedrock/agentc... Genesys Cloud Copilot: https://www.genesys.com/en-gb/copilots Thinking, Fast and Slow by Daniel Kahneman: https://www.amazon.com/Thinking-Fast-... Thanks for the Feedback by Douglas Stone & Sheila Heen: https://www.amazon.com/Thanks-Feedbac... Mythos by Stephen Fry: https://www.amazon.com/Mythos-Stephen... 👤 About the guests: John Sexton is a Manager on the Genesys Cloud Copilot team, with a background in software development and machine learning. Aaron Tummon is a developer at Genesys in Galway, Ireland, working on agentic AI after a career spanning embedded systems and automotive. 🎙️ About the AWS Developers Podcast: Weekly conversations about Agentic AI and the future of software development with customers, AWS community members, and industry luminaries. Hosted by Romain Jourdan.

Keynote: After the AI Hype – What’s Real, and What’s Next - Richard Campbell - 2026

Andrej Karpathy: From Vibe Coding to Agentic Engineering w/ Stephanie Zhan

Is AI Making Your Developer Experience Worse?

Turing Award Winner: Disagreeing with Google, Postgres, Future Problems | Mike Stonebraker

A leader’s guide to advanced team structures in an agentic world | AWS Events

Why your agent evaluations will fail you (and how to fix them before production)

Inside Anthropic, the $965 Billion AI Juggernaut | The Circuit

Is RAG Still Needed? Choosing the Best Approach for LLMs

Anthopic, OpenAI Should Not Be Allowed to IPO, Says Ed Zitron

Billionaire's WARNING: I'm SELLING. The Crash Is Already Here!

Dark factories: why your AI coding setup is already outdated

Cutting through the AI developer hype

Skill Issue: Andrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AI

Inside the Mind of Anthropic CEO Dario Amodei | The Circuit | Extended Interview

Creator of C++: Bell Labs, Negative Overhead Abstraction, Mistakes | Bjarne Stroustrup

Why AI Agents are either the best or worst thing we’ve ever built

Full Walkthrough: Workflow for AI Coding — Matt Pocock

5 quality gates that let you ship 250% faster with AI coding agents

AI Challenges in Software Development – REWORK

