Skip to content
AX Principles

Agent Experience · Part of uxuiprinciples

Design your product for the agents that now use it.

25 system-facing Agent Experience principles, each with a copy-paste artifact, a verifiable citation, and code. The reference for building agent-ready products.

01

Comprehensible

Expose your purpose in clean, machine-readable form so agents can use and cite it.

02

Accessible

Let agents authenticate with scoped identities and self-correcting errors, no human-only gates.

03

Actionable

Ship well-described, scoped tools with structured results agents can act on safely.

04

Accountable

Make agent activity visible, bound autonomy, and recover from failure gracefully.

Categories

Nine ways to handle agents

All categories →

Artifacts

Copy-paste, not theory

All artifacts →
# Acme
> One-line description of what Acme does.
## Docs
- [Quickstart](https://acme.com/docs/quickstart.md): get running in 5 minutes
- [API reference](https://acme.com/docs/api.md): endpoints and schemas
<script type="application/ld+json">
{"@context":"https://schema.org","@type":"SoftwareApplication","name":"Acme",
 "description":"Acme helps teams do X.","url":"https://acme.com",
 "offers":{"@type":"Offer","price":"29","priceCurrency":"USD"}}
</script>
GET /docs/quickstart HTTP/1.1
Accept: text/markdown
<link rel="alternate" type="text/markdown" href="/docs/quickstart.md">

How to use this

01

Read the principle

Each principle states what to do, why, and what to avoid, with a primary citation.

02

Copy the artifact

Grab the manifest, schema, error shape, or header and adapt it to your stack.

03

Ship and verify

Dogfood it: expose llms.txt, a capability manifest, and markdown twins like this site does.

Every principle cited.Bilingual EN / ES.Dogfoods llms.txt ↗Capability manifest ↗

Questions

AX is the discipline of designing products so AI agents can discover, authenticate to, act on, and recover within them, alongside the humans they serve.