Kore.ai puts Autoloop live to keep tuning enterprise agents

Kore.ai built Autoloop on two technologies it had already introduced: StateTrace, its evaluation layer, and the Agent Blueprint Language, which was brought out earlier this year and compiles routing, business rules and guardrails into an executable state machine. The pairing is what makes precise fixes possible — because each step in a trace maps back to a part of the blueprint, Autoloop can rewrite only the piece responsible for a miss. Kore.ai's chief technology officer and chief product officer, Prasanna Arikala, framed the logic bluntly: "You can't optimize what you can't see, or fix precisely what you can't express precisely."

The same approach runs inside Kore.ai itself, where AI agents write code. Roughly 6,500 commits a month to the company's 2.6 million-line production codebase now come from those agents, working under 68 always-on guardrails. Founder and Chief Executive Raj Koneru's read is that the companies able to scale AI "will be the ones using AI to build, govern and optimize AI."

The commercial bet looks aimed at the gap between deploying an agent and keeping it accurate. Kore.ai counts more than 500 Global 2000 organizations as customers, so Autoloop's spread on the Artemis edition may hinge on whether automatic tuning holds up once real interactions begin starting fresh rounds of optimization.