Salesforce today announced the launch of Headless 360, a fundamental architectural redesign of its core platform, during the opening keynote of the TrailblazerDX conference in San Francisco. The new framework is designed to decouple the Salesforce logic and data layers from the traditional graphical user interface, exposing the company’s entire suite of enterprise tools through a comprehensive set of application programming interfaces (APIs). This shift is intended to allow autonomous AI agents to perform complex business operations directly within the Salesforce ecosystem.
According to Salesforce Chair and CEO Marc Benioff, Headless 360 represents the third major evolution of the company’s technology stack, moving beyond the cloud and mobile eras into an agentic era. The architecture provides over 1,500 new atomic APIs that cover Sales, Service, Marketing, and Commerce functions. By utilizing these endpoints, AI agents can execute tasks such as updating opportunity stages, generating personalized marketing journeys, and resolving customer service inquiries without human intervention or the need to navigate a visual dashboard.
A key component of the announcement is the Agentic Orchestration Layer, a new management framework that provides governance for autonomous systems. Parker Harris, Salesforce Co-Founder and Chief Technology Officer, stated that this layer includes a Reasoning Engine that allows agents to plan multi-step workflows and an Action Library where administrators can define specific permissions. This ensures that AI agents operate within defined corporate guardrails and security policies. The system also integrates with the Einstein Trust Layer to provide an audit trail for every action taken by an AI agent, ensuring transparency in automated decision-making.
The Headless 360 architecture is built on top of Salesforce Data Cloud, utilizing a zero-copy data strategy. This allows the platform to ingest and act upon real-time data from external sources, including Snowflake, Amazon Redshift, and Google BigQuery, without the need for data duplication. Salesforce officials noted that this real-time data access is critical for AI agents to make accurate decisions based on the most current enterprise information.
Salesforce confirmed that Headless 360 will be rolled out in phases, with a developer preview beginning immediately. General availability for enterprise customers is scheduled for the third quarter of 2026. The company indicated that pricing for the new architecture would transition toward a consumption-based model, where costs are determined by the volume of API calls and successful agent-led transactions. This move aligns with Salesforce’s broader strategy to position its platform as the primary operating system for autonomous enterprise intelligence.