Platform Engineering for the Agentic Enterprise: When AI Agents Become Platform Users
A CNCF piece argues internal developer platforms must now treat AI agents as first-class consumers alongside human engineers — sharing one context, identity and governance model.
Internal developer platforms were built for people. A new CNCF article argues that assumption no longer holds: as AI agents start deploying components, changing configuration and responding to incidents, platforms must treat those agents as first-class consumers alongside human engineers — each accessing shared capabilities through the right interface.
Three first-class entities
The author reframes the enterprise software estate around three interdependent things: applications (business logic), resources (databases, APIs, storage) and AI agents (intelligent actors) — rather than treating resources as mere plumbing beneath the apps.
Context and one governance model
Making this work depends on shared context: a unified model connecting applications, resources, agents, deployment history, ownership and policy, so humans and AI can reason over the same system. The recommendation is not to spin up a separate 'AI platform' but to extend existing cloud-native platforms so identity, security, observability and policy apply consistently to every consumer.
At Cloud of Things, we help teams turn trends like this into practical next steps. Before you let agents act on your systems, make sure they inherit the same guardrails your people do.
Source: CNCF Blog — Lakmal Warusawithana (WSO2), 21 Jul 2026
