Kore.ai launches Autoloop optimization engine for AI agents

Kore.ai has announced the general availability of Autoloop, an optimization engine designed for its Agent Platform, Artemis edition. The tool aims to automate the building, measuring, and continuous improvement of enterprise AI agents based on specific business goals rather than manual prompt tuning.

Autoloop focuses on several key performance dimensions, including task completion, accuracy, grounding in enterprise data, adherence to business rules, and safety guardrails. The system is designed to optimize agents both before and after deployment by analyzing production interactions. A specific feature, StateTrace, allows for the tracing of full agent execution paths, including tool calls and handoffs.

According to Kore.ai CEO Raj Koneru, the engine is intended to help enterprises scale AI by ensuring agents follow rules and operate cost-effectively. The company reported that it already utilizes Autoloop internally to manage its own production codebase.