Salesforce on Thursday previewed a Trusted Enterprise AI Harness, a composable architecture that bundles six trusted capabilities with a new AI Control Plane so businesses can govern agents and models running across the enterprise from a single control layer.
Six trusted capabilities and one control plane
The Harness combines Trusted Context, Trusted Agency, Trusted Action, Trusted Governance, Trusted Security and Trusted Models on top of Salesforce's existing stack — Data 360, Informatica, MuleSoft, Agent Fabric, Tableau, Agentforce, Salesforce Guardian and the Salesforce Platform. The idea is to give agents a shared understanding of the customer and business (context), the ability to reason and plan (agency), a secure way to act on APIs and workflows (action), and consistent policy, security and model-routing controls across all of it.
Alongside the Harness, Salesforce is introducing an AI Control Plane where enterprises can discover and register agents and AI capabilities, establish identity and policy, manage lifecycle, evaluate performance, observe behavior and outcomes, and control cost — across both Salesforce and third-party AI.

Built for a world of 3.1 agent platforms per enterprise
Salesforce is betting that companies won't standardize on a single agent stack. VentureBeat's July 2026 Agentic Orchestration Pulse Survey found 85% of enterprises already run two or more agent orchestration platforms, with 3.1 platforms on average. The Harness is being built "headlessly" from the ground up and is accessible through MCP, APIs, Skills and Plug-ins — reaching agents and AI experiences that run inside Anthropic's Claude, Microsoft Teams, Slack, Agentforce and other surfaces as part of Salesforce's broader AIforce strategy.
Timing and availability
Many of the foundation technologies are available today. The unified experience and new capabilities are planned to begin rolling out in early fiscal 2028, with pricing, packaging and upgrade paths detailed closer to general availability. Rohan Kumar, Salesforce's president and chief platform and engineering officer, framed the pitch as "a system where every interaction and action can create new context and intelligence, making the enterprise's context richer over time."
Related coverage: Accenture's Gemini Enterprise push, Salesforce–Listen Labs talks and OpenAI's Agents API beta.
Reporting based on coverage from Salesforce Newsroom, VentureBeat and SiliconANGLE.
