LaunchDarkly launches AgentControl to manage AI agents in production
LaunchDarkly has released AgentControl, a system that lets engineering teams adjust how AI agents behave in production without redeploying applications. The tool can change an agent's runtime behaviour in under 200 milliseconds, including switching models or triggering fallback responses before users encounter problems.
The launch addresses a shift in how businesses operate AI. As organisations move agents from testing into customer-facing services, they face operational challenges that differ from traditional software releases. An agent's behaviour can change based on prompts, models, and production conditions even when the underlying code stays the same.
Managing fragmented AI operations
Once deployed, agent settings and management often scatter across multiple teams and frameworks. Engineering teams need shared rules for governance, version control, and release management - and the ability to intervene quickly as live conditions change.
AgentControl combines runtime intervention with broader operational oversight. It offers central visibility across teams and systems, benchmarking before changes reach customers, progressive roll-outs, trace-level monitoring, and the ability to adjust agents based on production data.
LaunchDarkly built its reputation in feature management and software release controls. The company applies that approach to AI systems, where intervention may happen during a conversation rather than over a standard deployment cycle.
Addressing inconsistency and drift
AI agents produce inconsistent outputs. Model performance shifts over time. Businesses using these systems need ways to monitor changes, test updates carefully, and reduce the risk of errors reaching end users.
Cameron Etezadi, Chief Technology Officer at LaunchDarkly, said the company recognised overlap between established software control problems and operational challenges emerging around AI agents.
"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," Etezadi said.
Industry backing for runtime controls
Cursor, which develops AI coding tools for software teams, has expressed support for the launch. Both companies are positioning runtime controls as essential infrastructure as organisations move from pilot projects to production deployments.
"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 of Global Revenue and Field Operations at Cursor.
LaunchDarkly serves thousands of customers worldwide, including a quarter of the Fortune 500, from its base in Oakland. For operations professionals managing AI systems, understanding AI for Operations and Generative AI and LLM governance is increasingly relevant to production responsibilities.
Your membership also unlocks:
