Salesforce has launched Salesforce Headless 360, a product that exposes core Salesforce functions as APIs, MCP tools and CLI commands for software agents and developers.
The launch reflects a broader shift to make more of the platform accessible outside the traditional Salesforce user interface.
Headless 360 brings together more than 60 new MCP tools, over 30 preconfigured coding skills and a new experience layer that lets agents interact with Salesforce services across channels including Slack, mobile and messaging apps. The goal is to let both people and software agents work from the same underlying data, workflows and governance controls.
For developers, Salesforce is positioning the launch as a way to use its services from coding environments already common across engineering teams. The tools target coding assistants such as Claude Code, Cursor, Codex and Windsurf, giving them access to data, workflows and business logic held inside Salesforce systems.
Salesforce also introduced Agentforce Vibes 2.0, which adds full organisation awareness from the outset and supports multiple models, including Claude Sonnet and GPT-5. Alongside it, DevOps Centre MCP is designed to bring the same programmatic access into CI/CD pipelines, using natural-language prompts to describe deployment tasks.
Salesforce says this can cut development cycle times by as much as 40% by combining tasks that previously required several tools into one workflow. Native React support is also being added for teams that want to design custom interfaces while continuing to use Salesforce as the underlying platform.
Elia Wallen, Chief Executive Officer, Engine, said the wider Agentforce platform had already been put into production quickly inside his company.
"With Agentforce, we've been able to deploy sophisticated, production-ready AI agents in just 12 days, driving millions in savings while significantly increasing our technical velocity. This unified platform proves that we can scale our most complex service needs without adding operational complexity," said Wallen.
Conversation Layer
A central part of the launch is the Agentforce Experience Layer, a new UI service that separates agent actions from how those actions are displayed. This allows rich components such as approval cards, workflow steps, decision tiles and data layouts to appear natively inside Slack and in other clients that support MCP apps, including ChatGPT, Claude, Gemini and Teams.
Salesforce argues this reflects a workplace trend in which conversations in collaboration tools are increasingly where work gets done. It says custom AI agents on Slack have grown by 300% since January, with Slackbot serving as a front-end entry point for many enterprise agent interactions.
Oliver Bodden, Senior Product Manager, Indeed, said the benefit was direct access to the company's Salesforce environment from tools engineers already use.
"At Indeed, our mission is to help people get jobs, and the faster we can innovate, the faster we can connect job seekers to the right opportunities and employers to the right talent. By building on Agentforce, we're able to give coding agents live access to our entire platform, directly within the tools we already use, so we can move from idea to implementation quickly. This paired with proper gating and human-in-the-loop best practices, the results are faster delivery, more consistent execution, and a much clearer path from experimentation to production impact," said Bodden.
Testing And Control
Salesforce also used the launch to outline a framework for governing AI agents both before release and after deployment. New products include Testing Centre, Custom Scoring Evals, Agent Script, Observability, Session Tracing and A/B Testing.
These tools are designed to help companies monitor probabilistic software systems whose outputs may vary from one interaction to another. Testing Centre is intended to identify logic gaps, policy breaches and inconsistent responses before users see them, while Custom Scoring Evals lets companies define what a correct outcome looks like for a given use case.
Agent Script is designed to let organisations specify which parts of an agent's behaviour must follow fixed business logic and which parts may rely on model reasoning. After deployment, Observability and Session Tracing are meant to show not just what happened during an interaction but why, while A/B Testing allows several agent versions to run against live traffic.
For businesses running agents across different vendors and platforms, Salesforce unveiled Agent Fabric, intended to provide centralised governance over agents, tools and language models. It is positioned as a single control plane for orchestration and oversight across a broader AI estate.
Platform Bet
Underlying the launch is Salesforce's argument that enterprise AI systems need more than model access and raw APIs. The differentiator, it says, is a platform that combines data context, established workflows, permissions and user-facing engagement channels.
Data 360 now exposes trusted business data through APIs, MCP tools and CLI commands, while Customer 360 provides workflows and business rules developed over years of enterprise use. Slack remains the engagement layer, and Agentforce is positioned as the management layer for building and running agents.
Adones Guerra, Tech Lead, Grupo Globo, said the company was already using Agentforce Vibes to automate repetitive development work.
"As a dev team lead, I've seen firsthand how Agentforce Vibes is supporting our Salesforce development. It's been especially valuable for speeding up day-to-day tasks like metadata updates and quick tweaks, reducing manual work and boosting efficiency. We primarily use it for repetitive, time-consuming work, from config analysis and metadata adjustments to generating boilerplate code, tests, and handling small refactors or pre-deployment checks. Vibes acts as a strong assistant that streamlines our workflows, and we're excited to see how it evolves and drives even greater impact over time," said Guerra.
Salesforce also said AgentExchange now combines 10,000 Salesforce apps, more than 2,600 Slack apps and more than 1,000 Agentforce agents, tools and MCP servers from partners including Google, Docusign and Notion. Separately, it has set up a $50 million Builders Fund to back developers and partners building around the ecosystem.