ServiceNow Charges AI Agents, SAP Blocks Them, DataDog Caps Access

Context Enterprise software vendors are confronting a surge of autonomous AI agents that can generate thousands of API calls without adding human seats. Traditional per‑user licensing no longer matches usage patterns, prompting firms to rethink monetisation and security.

Key Facts - At its Knowledge 2026 conference, ServiceNow introduced Action Fabric, an integration layer that all external AI agents must traverse to read data or trigger workflows. The company will meter each operation performed through this layer and bill customers accordingly. Anthropic’s Claude will be the first partner to use the connector. - In April, SAP revised its API policy to forbid third‑party AI agents from calling SAP APIs unless they run on SAP‑approved architectures. The rule blocks autonomous or generative AI that plans or executes API sequences without SAP’s sanction. SAP’s own Joule Agents remain permitted. - DataDog adopted a usage ceiling for its model‑context protocol server, the gateway that lets external AI agents interact with DataDog products. Customers may submit up to 5,000 requests per day or 50,000 per month, with exceptions possible.

What It Means ServiceNow’s per‑action pricing turns AI traffic into a new revenue stream, mirroring Workday’s recent shift toward consumption‑based billing. By charging for each AI‑driven operation, vendors can capture value from agents that would otherwise bypass seat‑based fees. SAP’s outright block forces partners to rebuild connectors on SAP‑approved stacks, steering integration revenue toward its own ecosystem. The move also signals a security‑first stance, limiting exposure to uncontrolled AI behaviour. DataDog’s caps preserve system stability while still offering AI‑enhanced insights. The limits create a predictable cost ceiling for customers but may push high‑volume users toward alternative platforms. Collectively, these strategies illustrate a split path: monetize AI activity, restrict it, or throttle it. Enterprises will need to monitor pricing models, compliance requirements, and request limits as AI agents become routine tools.

Looking ahead, watch how large customers negotiate usage thresholds and whether SaaS providers converge on a standard AI‑agent pricing framework.