Going Headless: Architecting for the Age of Agents

For thirty years, enterprise systems were built for one kind of user: a person clicking through a screen. Not anymore. The fastest-growing users of your systems are AI agents — and they interact with your backend directly, through open protocols like MCP. In this whiteboard session, MuleSoft's Jing Li traces how architecture evolved from monolith → microservices → headless, where any "head" — user, developer, or agent — can sit on top of your backend. She breaks down the two challenges of opening your systems to agents (structured access and governance) and the three capabilities that make an enterprise agent-ready without a rewrite. If you're preparing for the agentic era, this is the foundation you can't skip. About MuleSoft: Connect systems, data, and AI with MuleSoft, the integration and automation foundation for the agentic enterprise. As part of Salesforce, MuleSoft enables organizations to securely connect applications, APIs, and data so AI agents can discover capabilities, take action across systems, and operate with enterprise-grade governance and visibility. Follow us! • Facebook: https://www.facebook.com/mulesoft • Twitter: https://twitter.com/MuleSoft • LinkedIn: https://www.linkedin.com/company/mulesoft • Instagram: https://www.instagram.com/mulesoft Timestamps: 0:00 — Your Fastest-Growing Users Aren't People 0:35 — From Monolith to Headless 1:30 — The Three Heads: User, Developer, Agent 2:20 — The Two Challenges of the Agent Era 3:15 — Three Capabilities for Agent-Ready Headless 4:09 — Accessible, Discoverable, Governed — with MuleSoft Copyright © MuleSoft