Managing AI Agents at Scale with Google Cloud's Riyaz Habibbhai

Live from Google Cloud Next 2026 in Las Vegas, Sam Charrington sits down with Riyaz Habibbhai, Director of AI Product Marketing at Google Cloud, to unpack the biggest AI announcements from the event. The conversation focuses heavily on the shift from developer-centric tools to true enterprise-grade AI agents. Riyaz breaks down the launch of the Gemini Enterprise Agent Platform, explaining how Google is prioritizing "glass box" observability, governance, and seamless integrations using the Model Context Protocol (MCP). They also discuss the tangible business value of Google's "full stack" AI architecture—from TPUs to the application layer—and share real-world success stories, including how Macy's built a fully orchestrated voice agent in just four weeks. If your organization is trying to navigate the transition from AI hype to secure, scalable agentic workflows, this is the episode of the TWIML Briefing Room is for you! 🔔 Subscribe to TWIML AI Podcast for more deep dives into Machine Learning and AI: https://twimlai.com/ ⏱️ Chapters: 00:00 - Introduction: Live from Google Cloud Next 00:30 - Key Themes: The AI-Optimized Stack & Interoperability 02:35 - Why a "Full Stack" AI Approach Matters to Customers 04:40 - The Evolution of Vertex AI into the Gemini Enterprise Platform 07:00 - Bridging the Gap Between Developers and Knowledge Workers 08:30 - "Glass Box" Observability & Agent Governance (Evals, Traces, & IDs) 10:23 - Breaking Down Enterprise Silos with a Unified Platform 13:23 - Embracing Openness: The Role of MCP (Model Context Protocol) 16:01 - Real-World Agent Adoption & Customer Success (Macy’s Case Study) 18:40 - Advice for Enterprises: Start with a POC & Build the Business Case 19:23 - Conclusion 🔗 Links: Check out the TWIML Blog for more analysis: https://twimlai.com/articles Learn more about Gemini Enterprise Agent Platform: https://cloud.google.com/blog/products/ai-machine-learning/introducing-gemini-enterprise-agent-platform Follow Sam Charringto