Salesforce has launched the Trusted Enterprise AI Harness, a new enterprise-grade architecture aimed at enabling secure, scalable deployment of agentic AI across business functions. Announced on September 11, 2026, the platform integrates six foundational capabilities—context, agency, action, governance, security, and models—into a single, composable AI control plane that aligns with existing customer relationships, processes, and controls (salesforce.com).
As enterprises increasingly adopt AI agents that understand, decide, and act autonomously, Salesforce’s solution addresses a critical challenge: how to ensure these agents operate with trust and compliance at scale. The AI Harness provides a shared understanding of business context, enabling agents to reason and plan actions while adhering to enterprise governance and security policies (salesforce.com).
Salesforce plans to roll out the new capabilities starting in early fiscal year 2028, offering organizations a unified foundation rather than requiring them to build and manage these capabilities separately for each AI application. This move positions Salesforce as a key enabler for enterprises transitioning from pilot AI deployments to production-scale, agentic systems with robust oversight and control (salesforce.com).