Salesforce Unveils Enterprise AI Harness with Composable Architecture

Salesforce has unveiled its Enterprise AI Harness, a composable architecture designed to bring together trusted context, agency, action, governance, security, and models for AI to understand and operate across the enterprise. The introduction of this new platform is part of Salesforce's efforts to evolve its AI and data stack towards agents.

According to Rohan Kumar, recently appointed President and Chief Platform and Engineering Officer at Salesforce, the Enterprise AI Harness builds on existing technologies such as Headless 360, Data 360, Informatica, MuleSoft, Tableau, and Agentforce. The platform is designed to be durable as AI evolves, allowing enterprises to switch out models as new ones become available or old ones become expensive.

The six essential trusted capabilities of the Enterprise AI Harness are Trusted Context, Trusted Agency, Trusted Action, Trusted Governance, Trusted Security, and Trusted Models. Kumar notes that there are still gaps in governance and cost controls, but Salesforce is actively working on these aspects through research and development with customers.