Salesforce introduced its Enterprise AI Harness on September 10, formalizing six capabilities—Trusted Context, Trusted Agency, Trusted Action, Trusted Governance, Trusted Security and Trusted Models—under a common architecture for enterprise AI.
The announcement also introduces an AI Control Plane intended to give organizations one place to discover, register, govern and monitor AI agents and capabilities across Salesforce and third-party systems. The design is explicitly composable rather than limited to Salesforce-built models or agents.
The availability boundary matters. Salesforce says many underlying technologies already exist, but new Harness capabilities and the unified experience are planned to roll out sequentially beginning in early fiscal 2028. Pricing, packaging, upgrade paths and regional availability are to be announced closer to general availability.
Independent technical coverage confirms that the announcement largely unifies and formalizes existing Salesforce platform components rather than representing a fully deployed new stack today. Buyers should therefore separate the architecture Salesforce has announced from capabilities already available in production.
AiToolMap treats Salesforce's claims about business outcomes, security advantages and future operational benefits as vendor statements unless independently reproduced. The concrete current news is the architecture and control-plane roadmap; broad cross-vendor management at full scope remains a future availability claim.