AI-Native Cloud company DigitalOcean (NYSE: DOCN) announced on Monday that it has launched Agent Learning and Agent Data Services to help businesses develop self-improving AI agents, combining performance monitoring, testing and data management within its cloud platform.

Agent Learning Services comprises Insights, which records agent activity, Signals, which identifies failures and performance issues, and Evaluations and Simulations, which tests improvements before deployment. The enhanced Inference Router can select models based on cost, quality or customer-defined evaluations, with automated optimisation for agent workflows planned for a future release.

DigitalOcean's Agent Data Services provides a managed data layer that processes and indexes documents, video, audio, logs and agent sessions for semantic search and AI applications, with configurable redaction and access controls. Training and testing data can be drawn from the platform where customers explicitly opt in.

The company added that Agent Learning Services is available in public preview, while the Inference Router optimiser and Agent Data Services are in private preview. DigitalOcean plans to demonstrate the services at its Open Intelligence Summit in San Francisco on 13 October 2026.