LaunchDarkly has launched AgentControl, a new solution that gives software teams real-time control over AI agents in production. With AgentControl, teams can change how an agent behaves at runtime without redeploying the underlying application.

As AI agents move into production, engineering teams need new ways to manage configuration, quality, and runtime behavior. Unlike traditional code, agent behavior may vary across models, prompts, and production contexts, even when application code remains unchanged.

Once deployed, agent configurations can span teams and frameworks, creating a need for shared standards around governance, versioning, and release controls. Teams also need the ability to intervene quickly when production conditions change.

AgentControl was built to close that gap. Instead of separating visibility from lifecycle tooling, AgentControl combines runtime intervention with the operational capabilities teams need to run agents reliably: configuring agent behavior across teams and frameworks, benchmarking quality before changes reach production traffic, introducing updates through controlled rollouts, monitoring performance with trace-level visibility, and refining behavior based on production data, all without requiring redeployment.

That kind of control requires a different order of speed. With LaunchDarkly, configuration changes propagate in under 200 milliseconds, fast enough to change how an agent behaves, route to a different model, or trigger a fallback within the turn of a conversation, before a customer sees a bad response.

“LaunchDarkly has always been about giving software teams control at runtime over what their software does in production,” said Cameron Etezadi, CTO of LaunchDarkly. “The hardest problems in AI, like model drift, unpredictable outputs, and the inability to intervene fast enough, turn out to be exactly the problems our platform was built to solve. We didn’t have to reinvent the platform; we just had to extend it to meet the demands of an AI SDLC and agentic-driven workflows.”

“Cursor is how the world’s leading enterprises are building with AI. As more AI-powered products and agentic capabilities reach production, runtime control becomes essential infrastructure alongside the development workflows and controls teams already trust,” said Brian McCarthy, President, Global Revenue and Field Operations at Cursor.

“LaunchDarkly built an additional layer for that environment, and AgentControl extends it to the agent lifecycle in a way that complements how Cursor’s customers already build,” McCarthy concluded.