ARF AI

ARF AI provides governance infrastructure for teams using AI agents to make changes to production systems. Its control plane applies deterministic policy checks before a governed action is executed, helping organisations define which changes an agent may make, which require human approval, whether they can be undone, and what audit record is retained. The platform is focused first on storage, with demonstrations using a simulated ONTAP cluster.

ARF AI’s capabilities include policy enforcement, approval workflows, explainable risk scoring, cryptographically signed and tamper-evident audit trails, and operational monitoring. Its recorded demonstration shows how a proposed storage-volume deletion can be held for administrator approval, limited to the specifically approved action, and verified after execution. Actions the system cannot model are refused rather than forwarded.

The company serves platform and SRE leaders, technology founders, and teams adopting AI-driven operations, particularly those preparing to answer customers’ security and governance reviews. Offerings include a free simulation sandbox, a fixed-price review of an agent’s production write paths, and an invitation-based continuous enforcement option.