Salesforce launches control plane for enterprise AI
Salesforce is building a governance layer for companies running multiple AI agents, models and workflows.
As companies deploy more AI agents, managing those systems is becoming a software category of its own.
What happened
Salesforce unveiled its Trusted Enterprise AI Harness, an architecture combining context, agent actions, governance, security and model access. A new AI Control Plane is intended to let companies register agents, enforce policies, monitor behaviour and track costs across Salesforce and third-party AI systems.
Why it matters
Enterprises increasingly need centralised controls for AI in the same way they already manage identity, cloud infrastructure and security. Without that layer, agent deployments can become fragmented and difficult to govern.
The bigger picture
The enterprise AI stack is expanding from models and applications into management infrastructure. Control planes, observability and policy enforcement could become core software categories as agent adoption scales.