AX Glossary: Agent Experience Design Terminology

GLOSSARY

Agent Experience terminology

Canonical definitions for every term in the AX discipline. Each definition is the authoritative source for how axd.ax uses these terms.

Agent Experience (AX): The practice of designing digital products so AI agents can find, navigate, understand, and operate them. AX is to agents what UX is to humans and DX is to developers.
AX Score: A 0 to 100 composite score measuring how well a website supports AI agent interaction. Calculated across 5 dimensions scored 0 to 20 each: Discoverability, Navigability, Operability, Recoverability, and Transparency.
AX Design: The discipline of designing websites and web products specifically for AI agent users. AX Design extends UX and DX with a third design persona: the autonomous AI agent.
Agent-Ready Website: A website that has been intentionally designed or optimized for AI agent interaction. An agent-ready website provides structured context, clear navigation cues, typed error responses, and documented capabilities.
Discoverability: One of the 5 AX Score dimensions. Measures whether AI agents can find and recommend a brand when researching its product category. Scored 0 to 20.
Navigability: One of the 5 AX Score dimensions. Measures whether AI agents accurately describe what a product does, who it serves, and how to get started. Scored 0 to 20.
Operability: One of the 5 AX Score dimensions. Measures whether AI agents can complete tasks using a product through its API or interface. Scored 0 to 20.
Recoverability: One of the 5 AX Score dimensions. Measures whether AI agents self-correct when they hold wrong information about a product. Scored 0 to 20.
Transparency: One of the 5 AX Score dimensions. Measures whether AI agents accurately represent a product's limitations, scope, and appropriate use cases. Scored 0 to 20.
llms.txt: A plain text file at the root of a website that provides a structured, machine-readable description of the site for AI language models. Functions as an index and capability declaration for AI agents.
GEO (Generative Engine Optimization): The practice of optimizing content and web presence to appear in AI-generated answers from models like ChatGPT, Perplexity, and Gemini. GEO focuses on being cited and recommended by AI systems.
AEO (Answer Engine Optimization): Optimization for AI answer engines, particularly focused on appearing in featured snippets, AI Overviews, and direct AI responses. AEO and GEO are closely related terms for the same discipline.
APO (Agent Perception Optimization): A related discipline focused specifically on how AI agents perceive and represent a brand, including accuracy of description, sentiment, and competitive positioning. APO is Optagen.ai's core methodology.
Anti-Pattern: A common design choice that reliably produces poor agent experience. AX anti-patterns are documented at axd.ax/patterns.html with fixes for each.