Salesforce Headless 360: The CRM You Never Log Into

Salesforce just pulled off the biggest architectural transformation in its 27‑year history. At TDX in San Francisco, the company announced Headless 360 — a sweeping initiative that exposes every capability in its platform as an API, MCP tool, or CLI command so AI agents can operate the entire system without ever opening a browser. Over 100 new tools and skills shipped immediately. The message from Marc Benioff was blunt: "No browser required. Our API is the UI." This isn't a minor API refresh. It's Salesforce answering the existential question hanging over enterprise software: in a world where AI agents can reason, plan, and execute, does anyone still need a CRM with a graphical interface? Salesforce's answer is no — and that's the whole point. As the company put it: "We made a decision two and a half years ago: Rebuild Salesforce for agents. Instead of burying capabilities behind a UI, expose them so the entire platform will be programmable and accessible from anywhere."

Headless 360 rests on three pillars. The first — build any way you want — delivers more than 60 new MCP (Model Context Protocol) tools and 30+ preconfigured coding skills that give external agents like Claude Code, Cursor, Codex, and Windsurf complete, live access to a customer's Salesforce org, as VentureBeat reported. Developers can direct AI coding agents from any terminal to build, deploy, and manage Salesforce applications — no Salesforce IDE required. Agentforce Vibes 2.0 now includes an "open agent harness" supporting both the Anthropic agent SDK and the OpenAI agents SDK, letting developers switch between Claude Code and OpenAI agents depending on the task. And in a major break from tradition, Salesforce now supports native React — not just its own Lightning framework — connecting to org metadata via GraphQL while inheriting all platform security primitives. The second pillar — deploy on any surface — centers on the new Agentforce Experience Layer. This separates what an agent does from how it appears, rendering rich interactive components natively across Slack, mobile apps, Microsoft Teams, ChatGPT, Claude, Gemini, and any MCP‑compatible client. As CX Today reported, developers define a workflow once and deploy it across six different surfaces without writing surface‑specific code. The philosophical shift: rather than pulling customers into a Salesforce UI, enterprises push branded, interactive agent experiences into whatever workspace their customers already inhabit. The third pillar — build agents you can trust at scale — introduces lifecycle management tools spanning testing, evaluation, experimentation, observation, and orchestration. Agent Script, a new domain‑specific language for defining agent behavior deterministically, is now generally available and open‑sourced. A new Testing Center surfaces logic gaps and policy violations before deployment. Custom Scoring Evals let enterprises define what "good" looks like for their use case. And a new A/B Testing API enables running multiple agent versions against real traffic simultaneously.

The most candid portion of Salesforce EVP Jayesh Govindarjan's interview with VentureBeat addressed the fundamental tension at the heart of enterprise AI: agents are probabilistic systems, but enterprises demand deterministic outcomes. Early Agentforce customers discovered a painful reality after getting agents into production. "They were afraid to make changes to these agents, because the whole system was brittle," Govindarjan told VentureBeat. "You make one change and you don't know whether it's going to work 100% of the time. All the testing you did needs to be redone." Agent Script solves this by functioning as a single flat file — versionable, auditable — that defines a state machine governing how an agent behaves. Enterprises specify which steps must follow explicit business logic and which can reason freely using LLM capabilities. It's the anti‑vibe‑coding: deterministic structure wrapping probabilistic reasoning. Claude Code can already generate Agent Script natively because of its clean documentation. Govindarjan also drew a revealing distinction between two fundamentally different agentic architectures. Customer‑facing agents demand tight deterministic control — Agent Script encodes these as a static graph, a defined funnel of steps with LLM reasoning embedded within each step. The opposite is what he called the "Ralph Wiggum loop" — a dynamic graph that unrolls at runtime, where the agent autonomously decides its next step based on what it learned previously. This shows up in employee‑facing scenarios: developers using coding agents, salespeople running research loops. The critical insight: both architectures run on the same underlying platform and graph engine. "It's all a graph underneath."

Salesforce's embrace of openness at TDX was striking — the platform now integrates with OpenAI, Anthropic, Google Gemini, Meta's LLaMA, and Mistral AI. The new AgentExchange marketplace unifies 10,000 Salesforce apps, 2,600+ Slack apps, and 1,000+ Agentforce agents, tools, and MCP servers from partners including Google, Docusign, and Notion, backed by a $50 million AgentExchange Builders Initiative. But Govindarjan offered a candid assessment of MCP itself: "To be very honest, not at all sure" that MCP will remain the standard. "When MCP first came along as a protocol, a lot of us engineers felt that it was a wrapper on top of a really well‑written CLI — which now it is. A lot of people are saying that maybe CLI is just as good, if not better." The Headless 360 naming itself reflects this hedging: rather than betting on a single protocol, Salesforce exposes every capability across API, CLI, and MCP, insulating itself against protocol shifts. Underpinning all of it is a shift in how Salesforce gets paid. The company is moving from per‑seat licensing to consumption‑based pricing for Agentforce — a transition Govindarjan described as "a business model change and innovation for us." When agents do the work, per‑user pricing no longer makes sense. The timing is sharp: the iShares Expanded Tech‑Software Sector ETF is down roughly 28% from its September peak, driven by fear that AI could render traditional SaaS business models obsolete.

Headless 360 matters for three reasons. First, if you build tools that integrate with Salesforce (and most B2B tools do), the integration surface just exploded — 60+ MCP tools and 30+ coding skills mean you can now build agent‑native integrations that operate entirely through APIs and CLI, not screen‑scraping or brittle UI workflows. Second, the Agent Script open‑source release gives you a template for solving the agent brittleness problem that plagues every production AI deployment. The deterministic‑state‑machine‑meets‑probabilistic‑reasoning pattern is replicable beyond Salesforce. Third, Salesforce's move from per‑seat to consumption‑based pricing signals where the entire SaaS industry is heading. If your pricing model assumes human users, start planning for a world where agents are the primary operators. The proof points are already here: Engine, a B2B travel company, built its customer service agent Ava in 12 days and now handles 50% of cases autonomously. SaaStr's Jason Lemkin reported his team has been running headless Salesforce for six months — their AI VP of Marketing runs their entire GTM motion on top of Salesforce APIs, assigning tasks to humans and pulling real‑time pipeline data autonomously. As Salesforce co‑founder Parker Harris asked last month: "Why should you ever log into Salesforce again?" If Headless 360 works as designed, you shouldn't have to.