Salesforce has just unveiled "Headless 360," a monumental shift in its 27-year history, aiming to transform its entire platform into an infrastructure accessible by AI agents without needing a graphical interface. This initiative exposes every Salesforce capability as an API, MCP tool, or CLI command, allowing AI agents to fully operate the system. The move comes as a direct response to the rising influence of AI and large language models, addressing the industry's concern that traditional SaaS models might become obsolete.
At the heart of Headless 360 are three core pillars. The first, "build any way you want," provides over 60 new MCP tools and 30+ co ding skills, granting external AI coding agents like Claude Code and Cursor complete access to a customer's Salesforce organisation. Developers can now direct these agents from any terminal to build and manage Salesforce applications. The platform also now supports native React, offering greater front-end flexibility beyond Salesforce's own Lightning framework. The second pillar, "deploy on any surface," introduces the Agentforce Experience Layer, decoupling agent actions from their presentation and enabling rich interactive components to be deployed natively across various platforms like Slack, mobile apps, and even ChatGPT. This means enterprises can push branded agent experiences directly into their customers' preferred workspaces. The third pillar, "build agents you can trust at scale," focuses on robust lifecycle management tools, including a new testing centre, custom scoring evaluations, and an A/B testing API, all designed to ensure AI agents function reliably and determinis tically.
A key technical innovation is "Agent Script," a new, open-sourced domain-specific language that merges the determinism of traditional programming with the flexibility of LLMs. This language helps manage agent behaviour, ensuring predictable outcomes crucial for enterprises. Salesforce has also differentiated between two agent architectures: tightly controlled, static graphs for customer-facing agents, and dynamic, "Ralph Wiggum loops" for employee-facing tasks where experts review the output. Both architectures run on the same unified platform. Salesforce is embracing openness, integrating with major AI models from OpenAI, Anthropic, and Google, while also hedging its bets on protocols like MCP by offering API, CLI, and MCP access. The company is shifting its business model from per-seat licensing to consumption-based pricing, acknowledging that AI agents are now performing the work. Headless 360 represents a fundamental transformation, making the entire Salesforce pl atform programmable and agent-friendly, betting that its accumulated enterprise data and logic will provide lasting value in the AI era.
Fuente Original: https://venturebeat.com/technology/salesforce-launches-headless-360-to-turn-its-entire-platform-into-infrastructure-for-ai-agents
Artículos relacionados de LaRebelión:
- Claude Managed Agents Solucion Unica o Riesgo
- AI Agents Zero Trust Architectures Secure Credentials
- AI Agents 5 Frameworks 3 Critical Security Gaps Exposed
- Cloudflares Dynamic Workers Run AI Agents 100x Faster
- Nvidias Vera Rubin Revolutionary Seven-Chip AI Platform
Artículo generado mediante LaRebelionBOT
No hay comentarios:
Publicar un comentario