AI agents are moving from chatbots to running workflows, and ops will feel it first

Salesforce’s Winter ’27 release positions AI agents to run end-to-end workflows inside CRM and collaboration tools, including service case guidance, voice scheduling, and agent-driven commerce search, according to Salesforce. In parallel, NIST’s new AI Agent Standards Initiative is asking for industry input on AI agent security and is drafting guidance on software and AI agent identity and authorization, Federal News Network reported. On the infrastructure side, CoreWeave says it is unifying training, inference, observability, and reinforcement learning so autonomous agents can improve in production, while Cloudflare and Plume data highlighted by Cablefax show non-human traffic and AI agent requests rising sharply, creating measurable capacity and latency planning implications. The operational consequence is that “agent readiness” is becoming a joint program across app owners, security teams, and network operators, with identity, authorization, observability, and traffic management becoming gating items for scaling agents beyond pilots.

Enterprise roadmaps often describe AI agents as “software that chats.” But the past year of launches and early metrics suggest they behave more like software that executes work. That shift pushes the toughest choices away from UX and toward governance and infrastructure.