LS-Labs

Glossary

What Is Agent Experience (AX)? Definition and Four Areas

Mathias Biilmann's term, organized by his four areas: access, context, tools and orchestration.

What is agent experience (AX)?

AX, or agent experience, is the overall experience an AI agent has when it is the user of a product or platform.

The term was introduced in January 2025 by Mathias Biilmann, CEO and co-founder of Netlify.

The four areas of AX

In One Year of AX, published in January 2026, Biilmann listed four areas that he thinks any digital product should start investing in. The questions below paraphrase his.

Biilmann's four areas of agent experience
AreaQuestions to ask
AccessCan the agent get into your product, does it have the right permissions, and must a human be in the loop? On a website, that means whether agents can get in at all, and whether steps such as payment or confirmation should be handed back to a person.
ContextDoes the model know about your product, does it have the right context to use it, and do your responses help it build the right context? Biilmann's examples include MCP, Skills, and content negotiation that makes docs sites return pure Markdown to agents.
ToolsDo you offer the right tools for agents, and can you reduce friction for a person working through an agent? Tools can be provided through MCP or Skills, or agents can call your CLI or API directly.
OrchestrationCan you trigger agent runs from your product, pass along the right context, and offer sandboxes and environments?

From UX and DX to AX

Biilmann sets AX alongside two earlier terms, UX and DX. Don Norman coined UX in 1993 to cover all aspects of a person's experience with a system. Jeremiah Lee coined DX in 2011 for the experience developers have building on top of a platform. AX puts an AI agent in the user's seat. Biilmann also expects AX to go from being understood mainly as a new kind of developer experience to being as generally applicable as UX.

How AX relates to MCAC and TSO

Compared with other terms on this site: MCAC separately checks brand mentions, source citations, factual accuracy and conversion-readiness in AI answers. TSO looks at whether an AI agent finds and selects your tool. AX looks at whether an agent can get the job done on your product. This is LS-Labs' simplified division of labor. When Biilmann discusses access, he also counts the need for agents to know about your product, and the emerging field of GEO, so his AX covers more ground.

Scope and limits

AX was introduced by Biilmann and is not an LS-Labs framework; this site uses his definition and four areas. AX is also not a ranking factor published by any search engine or AI platform.

Evaluate operation, delegated authorization and result evidence separately. One successful run or a stored record cannot replace verification of the complete task.

The four areas do not need to be tackled all at once. When Biilmann expands on orchestration, he singles out those building platforms that developers rely on. A typical brand website can start with access, context and tools.

Sources and scope

Sources as checked for the source-language entry on 2026-10-27; this English edition adds no new claims.

← Back to glossary