LaunchDarkly launches Release Autopilot for AI code releases

LaunchDarkly has launched Release Autopilot, automation intended to help engineering teams ship more AI-written code while reducing manual release work and automating safeguards against production issues. The launch comes as AI speeds code creation, which has made manual release processes an increasingly significant bottleneck in software delivery.

LaunchDarkly’s own pull requests reflect the shift, with AI assisting on more than 80% of them while merged pull requests grew 2.1 times with roughly the same number of engineers. Its 2026 AI Control Gap Report found that 94% of respondents said AI has increased their code velocity, while 91% said it has made them more cautious about pushing to production.

Release Autopilot assesses each change’s type, impact, and risk, and places changes that warrant control behind a feature flag. It then exposes them to progressively larger audiences while monitoring error rates, latency, and other release metrics. If a change breaches a threshold, the tool pauses the rollout or returns the change to its last good state without a redeploy, and it removes temporary flags once rollout completes.

Engineers set the policies and decide how much automation to use, ranging from reviewing and approving the tool’s recommendations to letting it act on its own. Release Autopilot operates alongside the agents, pipelines, and clouds that form a company’s software factory and runs on the LaunchDarkly platform, which evaluates 70 trillion feature flags a day.

“Human attention is among the scarcest resources in the software factory, and it belongs on the highest-value work,” said Jonathan Nolen, chief product officer of LaunchDarkly. Rachel Stephens, Research Director at RedMonk, said, “Release Autopilot is a clear step toward making progressive delivery a repeatable, low-friction default rather than an occasional risk-mitigation tactic.”

Release Autopilot will be generally available Nov. 16, with some features in beta.