What is Agent Experience (AX)?
The step after UX: designing sites, APIs and products so AI agents can use them too. Block them and you vanish from their flows.
Agent Experience (AX) is the discipline of designing digital products for a new kind of user: AI agents acting on people's behalf. The term was launched in early 2025 by Netlify's CEO as the natural evolution of UX: if a growing share of visits, purchases and integrations flows through agents, the experience you offer them becomes as important as the one you offer humans. Unlike classic UX, which is measured through usability tests on real people, AX is judged by whether an agent can read a page, understand a form, or complete a checkout without tripping over a captcha or a layout built only for human eyes. It is still a young discipline, but it is already showing up in briefs for e-commerce sites, B2B portals and technical documentation meant to be consumed by AI assistants as much as by people.
What it means concretely
An agent-friendly site or product has content readable without executing JavaScript, semantic HTML and structured data (the same things SEO and GEO reward), documentation in formats a model digests well, APIs with authentication designed for automated flows too, and forms and processes that do not depend on captchas or visual steps impossible for a machine. The touchpoints with the agent world are concrete: protocols like MCP and files like llms.txt are the first bricks of this infrastructure.
Why it matters now
Because the selection is already happening: when an agent compares suppliers, fills an order or looks for an answer, the sites that block it simply drop out of the flow, with no human noticing. For anyone building digital products the new design-phase question is: "could an agent complete this journey?". It is why in my web projects I treat technical SEO, GEO and agent readability as requirements from the first commit, not as finishing touches.
Related terms
- GEO & AEO · The practices for getting cited in AI answers (ChatGPT, Perplexity, AI Overviews): SEO's evolution in the era of generative engines.
- AI Agents (Agentic AI) · AI systems that go beyond answering: they plan, use tools and take actions autonomously inside your processes.
- Agent harness · The software scaffolding around an LLM that makes it an agent: the execution loop, the tools, the context, the limits.
- Generative UI · The interface a model assembles at runtime, not prebuilt in advance: distinct from personalization and adaptive UI.
- WebMCP · W3C proposal letting a website expose its own actions as tools an AI agent in the browser can invoke directly.
A term that hits close to home? Let's talk.
CONTACT ME