Agent experience (AX)
Also called AX, agentic experience
Agent experience is how well an AI agent can use a product on someone’s behalf: find what it needs, understand the page, and finish the task without a person stepping in.
An agent reads a page differently from a person. It works from the markup, the accessible names on controls and whatever text is on screen, and it cannot guess its way past a button labelled with an icon or a form that only works with a mouse.
Much of what makes a page easy for an agent is familiar from accessibility and technical SEO: meaningful labels, predictable navigation, forms that can be filled by keyboard, and content that exists in the page rather than only in an image. Some of it is newer, such as an llms.txt file or an MCP server that lets an agent act without driving the interface at all.
Agent experience matters once agents start buying, booking and comparing for people. A step an agent cannot complete is a sale the person never sees.
Where Stunt Double fits
An actor is itself an agent in a real browser, so a checklist or interview run shows you the exact step where an agent stalls, with the screenshot. The Stunt Double Index scores real websites on the same question in public.
Related terms
- Browser agentA browser agent is an AI model that operates a web browser: it reads the page, decides on an action such as a click or a keystroke, performs it, and repeats until the task is done.
- llms.txtllms.txt is a plain-text Markdown file at the root of a website that tells language models what the site is and links to the pages worth reading, in a form that is cheap to fetch and parse.
- Model Context Protocol (MCP)The Model Context Protocol is an open standard for connecting AI assistants to tools and data. A service exposes an MCP server, and any MCP client, such as Claude or Cursor, can call its tools.
- Synthetic usersSynthetic users are AI agents given a persona (a background, goals and constraints) that stand in for real users in research or testing.