Controlling Code and Agents in the AI Era
About this webinar
Software is increasingly generated, modified, and operated by machines. AI accelerates how quickly software changes and agents introduce behavior that shifts after deployment. Change is constant—and already live.
The systems we built to manage software assumed a slower world: code review, staging, deploy, then monitor. But that model breaks when behavior can change without a release, and failures show up in real time.
Join LaunchDarkly CEO & Co-founder Edith Harbaugh, CTO Cameron Etezadi, and Head of AI Marek Poliks for a practical discussion on what comes next—and what modern engineering teams need to safely operate AI systems at scale. The session will also introduce AgentControl, LaunchDarkly’s new solution for controlling and governing AI agents in production.
- We’ll discuss runtime control, a model where systems are continuously evaluated in production, risk is constrained as changes happen, and the system automatically responds to issues before users are impacted.
- We’ll demo AgentControl, a new LaunchDarkly solution for configuring, evaluating, observing, and controlling AI agents in production.
- And we’ll share how we’re applying these ideas at LaunchDarkly as we evolve our own development process with AI—including what’s worked, what hasn’t, and what we’re learning along the way.
In this era, speed isn’t the risk. Lack of control is.
