Salesforce unveiled “Headless 360,” its most ambitious architectural transformation in 27 years, aiming to make every capability in its platform accessible via APIs, MCP tools, or CLI commands for AI agents. This initiative arrives in response to significant concerns in the enterprise software sector about the potential obsolescence of traditional SaaS models in light of AI advancements.
The company introduced over 100 new tools and skills for developers as part of this initiative at its TDX developer conference. Salesforce’s decision to overhaul its platform for AI integration began two and a half years ago, emphasizing the need for a system that operates without the constraints of a graphical user interface.
Jayesh Govindarjan, EVP of Salesforce, highlighted that the transformation is informed by lessons learned from deploying AI agents with enterprise customers. “We made a decision two and a half years ago: Rebuild Salesforce for agents,” Govindarjan said. This reformation seeks to expose Salesforce’s full capabilities to make the platform programmable and accessible from any point.
Headless 360 consists of three main pillars. The first, “build any way you want,” provides over 60 new Model Context Protocol (MCP) tools and more than 30 preconfigured coding skills, granting external coding agents full access to the Salesforce platform’s functionalities. Developers can now create applications directly from any terminal without needing Salesforce’s IDE.
The second pillar, “deploy on any surface,” introduces the Agentforce Experience Layer, which enables agents to function across multiple platforms—including Slack, Microsoft Teams, and ChatGPT—without specific coding for each surface. Presenters demonstrated deploying an experience across six surfaces without writing dedicated code for each interface.
The third pillar, “build agents you can trust at scale,” features new lifecycle management tools that include testing, evaluation, and orchestration. A new open-source programming language, Agent Script, is now available, designed to define agent behavior deterministically and enhance system stability.
Salesforce is also shifting its pricing model from per-seat licenses to a consumption-based model for Agentforce. This adjustment reflects the company’s recognition that as agents perform the bulk of the work, traditional user-based pricing is no longer applicable.
Salesforce’s open ecosystem strategy supports integration with major AI models from OpenAI, Anthropic, and Google Gemini. The introduction of the AgentExchange marketplace unifies thousands of applications and frameworks underpinned by a $50 million initiative to support builders.
The company emphasizes that the AI agents designed to potentially supplant CRMs are now being utilized to enhance its own capabilities. “The software industry is still alive, well and growing,” Salesforce co-founder Marc Benioff remarked previously, positioning Headless 360 as a critical effort to assert Salesforce’s relevance in a rapidly evolving technological landscape.
Parker Harris, Salesforce’s co-founder, succinctly questioned, “Why should you ever log into Salesforce again?” indicating that the platform’s future may depend on user accessibility without traditional login requirements, fostering a new operational paradigm for enterprise solutions.
