您的浏览器禁用了JavaScript(一种计算机语言,用以实现您与网页的交互),请解除该禁用,或者联系我们。 [PitchBook]:无头、自主、不可阻挡:Salesforce再次做到这一点(英)2026 - 发现报告

无头、自主、不可阻挡:Salesforce再次做到这一点(英)2026

信息技术 2026-05-06 PitchBook 有梦想的人不睡觉
报告封面

EMERGING TECH RESEARCHHeadless, Autonomous, and Institutional Research Group Unstoppable: SalesforceDoes It Again Rudy TorrijosDirector, Industry Researchrudy.torrijos@pitchbook.com pbinstitutionalresearch@pitchbook.com Published on April 21, 2026 The absolute need for AgentOps to govern enterprise AI Contents PitchBook is a Morningstar company providing the most comprehensive, mostaccurate, and hard-to-find data for professionals doing business in the private markets. Key takeaways •Machine UX is born:Salesforce’s Headless 360 marks the inflection point forthe enterprise-grade availability of AI agents operating natively via APIs andmodel context protocols (MCP). The reliance on visual dashboards is removed toaccelerate agentic AI adoption. •Software is merging with the human labor market:Systems of record aretransforming into invisible, programmable systems of action. This prompts afundamental shift in the workforce, transitioning human capital from executingindividual tasks to acting as managers of automated workflows. •The “autonomy gap” triggers an AgentOps gold rush:Removing the visual safetynet of human oversight introduces significant operational risk. This drives massiveinvestment in AgentOps infrastructure to safely manage and govern the digitalworkforce, such as session tracing, decision making audit trails, and deterministicguardrails. •Business models must pivot to outcome-based pricing:The traditional seat-basedsoftware subscription model is obsolete for agentic applications. Venture-backed startups, PE sponsors, and legacy SaaS platforms must pivot tocapture incremental enterprise labor budgets by offering automated workflowservice pricing. •Valuation premiums will be driven by data moats and compliance:The highestM&A premiums will go to assets possessing proprietary “ground truth” data andtrusted business logic, self-correcting feedback loops, and pre-cleared regulatorymaster service agreements (for example, FedRAMP) that act as compliantdistribution channels for larger AI orchestrators. Executive summary The enterprise software industry has crossed a permanent threshold. Salesforce’srollout of its Headless 360 architecture marks the defining pivot to “service as software”(SaS) we previously described in ourSaaS is DeadandThrough the Looking Glassresearch. By eliminating the graphical user interface (GUI) from its products, Salesforceis acknowledging that the future of enterprise relies on AI agents operating natively viaAPIs, model context protocols (MCP), and command-line interfaces (CLI). The dashboardis officially a legacy artifact, and over time, no browser will be required. Salesforce’s aggressive move validates that the transition from a passive software modelto an active intelligence partner is the only path forward for enterprise incumbents.Though it is counterintuitive, the software market has begun to merge with the humanlabor market, transforming systems of record like Salesforce into a data and functionalitylayer that allows both Salesforce and third parties to build invisible, programmablesystems of action on top. This allows human capital to transform from task workers toaccelerated human platforms capable of orchestrating fleets of agents. The removalof the GUI, however, eliminates the visual safety net of human oversight. Consequently,the rapid deployment of autonomous AI demands an unprecedented acceleration inAgentOps innovation to manage, secure, and govern the digital workforce. Implications: catalysts, pivots, and premiums Below is a breakdown of the catalysts driving adoption, the necessary business modelpivots, and the assets that will command premiums. What Headless 360 achieves For 20 years, the constraints of software were defined by human boundaries. Thefundamental economic unit was a license granting an employee access to a visualinterface. Headless 360 fundamentally breaks this paradigm by decoupling the back-end data and deep business logic from the front-end experience. When a third-party AI agent attempts to navigate a traditional SaaS interface, the GUIlayer creates friction as the agent needs to virtually click on the correct action button.Headless 360 removes this entirely, exposing Salesforce’s underlying data graph anddeterministic rules directly to the agent orchestration layer. •Native agent access:The introduction of over 60 new MCP tools and preconfiguredcoding skills gives external agents—such as Claude Code, Codex, or Cursor—live,authenticated access to the platform without ever touching a browser, eliminatingUI obstacles. •The experience layer:This new architecture separates an agent's computationalaction from its graphical rendering. A workflow generated by an agent can now bedisplayed natively as an approval pop-up inside Slack, Microsoft Teams, or a voiceinterface, allowing continuous workflow from messaging apps without forcing auser into a CRM dashboard. •Agent script open-sourcing:By formalizing its definition langua