Learning Services combines three parts. Insights records agent activity, Signals flags failures and performance gaps, and Simulations tests possible improvements through realistic simulated conversations before deployment.
A separately described enhanced Inference Router also plays a role in the loop. It selects models based on task requirements, weighs cost against quality, and reruns the same evaluations when new model versions appear.
Agent Data Services is a managed data layer that processes and indexes documents, video, audio, logs and agent sessions, making them searchable through a catalog with governance and access controls. Optimizer draws training and test data from it only when the customer opts in. According to the article, Insights, Signals and Evaluations including Simulations are in public preview, while Inference Router optimization and Agent Data Services are in private preview.
Why it matters if you build agents
Teams running agents can record sessions, find failing tool calls and test fixes before rollout without building that pipeline themselves.
Anyone feeding agent traces into evaluation or training should review the customer opt-in rules, since the data layer is still in private preview.
Get stories like this every weekday
One short email with the day's agent tooling news.