Putting AI to work: governing enterprise AI agents (preview)

2026 is being called the year AI gets to work, but moving autonomous agents from pilots to production is where most enterprises hit a wall: once agents have read-write access to sensitive systems, the challenge shifts from "can we build this" to "can we control it." In this Tech Talk, Redpanda CTO Tyler Akidau and Principal Engineer Tyler Rockwood break down what it takes to run agentic systems safely at scale: - Why agentic workflows drift, improvise, and fail in ways that are hard to observe or recover from - How a centralized access layer between agents and enterprise systems changes the risk profile - What unified authentication, fine-grained authorization, and replayable agent transcripts look like in practice Watch the full session on AI agent governance at scale: https://www.redpanda.com/events/adp-318