Gabe Monroy of Workday, chief technology officer, joins theCUBE Research with host John Furrier to examine how agentic artificial intelligence moves from demonstrations to production. Monroy outlines the technical interplay between probabilistic models and authoritative systems of record and demonstrates integrations between Gemini Enterprise and Workday. They describe patterns for embedding agents into Google Workspace workflows while preserving operational controls and auditability. The discussion highlights the necessity of marrying probabilistic reasoning with deterministic guardrails and solving this at the inference-engine layer. Monroy emphasizes curating tool surfaces to prevent tool confusion and enforcing identity and authorization and auditability when agents access systems of record. The conversation also covers protocols and sandboxing such as Agent2Agent, Agent Substrate and OpenShell as deployment patterns for secure auditable enterprise agents. The episode examines enterprise AI adoption considerations including governance, observability and secure deployment in cloud-native environments.