Kore.ai's AI agent optimization engine Autoloop available

Kore.ai launched Autoloop, an optimization engine for AI agents that continuously builds, evaluates, diagnoses and improves them to meet business goals.

Autoloop, which is generally available, is designed to track task completion, accuracy, safety, customer experience and costs. Autoloop is part of the Kore.ai Agent Platform, Artemis edition, which launched in May.

Kore.ai's launch is the latest effort to up the AI agent observability game and optimize performance. Kore.ai CEO Raj Koneru said Autoloop is designed to improve the scaling of AI agents. "Every enterprise knows what it wants from its agents: finish the job, follow the rules, stay safe, and do it at a sensible cost," said Koneru.

• Constellation ShortList™ Agentic Platforms of the Enterprise

Autoloop is designed to start from goals like task completion, accuracy, business rule adherence and consistency with guardrails.

Here's how it works:

Key points about Autoloop include:

• The system optimizes agents before and after deployment based on production interactions.
• StateTrace, a feature in Autoloop, traces the full agent execution path including every handoff, delegation and tool call.
• Agent Blueprint Language (ABL) enables Autoloop to change problematic components by compiling various steps into one machine.
• Kore.ai uses Autoloop internally for its own AI agents, which produce about 6,500 commits a month on a production codebase of 2.6 million lines.

The launch of Autoloop is part of Kore.ai's strategy to be the AI harness focused on optimizing AI agents with governance.