Agentic Experience Design (AXD) is the discipline for designing trust-governed relationships between humans and autonomous AI systems. Founded in September 2024 by Tony Wood in Manchester, United Kingdom, AXD addresses how humans delegate, calibrate, observe, interrupt, and recover trust in agentic AI.
| Dimension | Traditional UX | Agentic Experience Design (AXD) |
|---|---|---|
| Primary material | Attention and affordance | Trust and delegation |
| User state | Present, navigating | Absent, delegating |
| Design output | Screens and interfaces | Outcomes and constraints |
| Temporal model | Session-based | Relationship-based |
| Success metric | Task completion | Trust calibration |
Agent-ready describes an organisation that has structured its products, services, data, and operating model so that autonomous AI agents can discover, evaluate, and transact with it reliably. It encompasses machine-readable content, verifiable trust signals, clear authority models, programmatic transaction capability, and designed recovery paths.
Businesses prepare by addressing four pillars: Signal Clarity (making offerings machine-discoverable), Reputation via Reliability (providing verifiable trust signals), Intent Translation (aligning value propositions with agent query patterns), and Engagement Architecture (enabling end-to-end programmatic transactions). The AXD Readiness Assessment provides a structured evaluation framework.
Agentic commerce requires structured data and schema markup for discoverability, verifiable performance metrics for trust, API-first transaction surfaces for execution, delegation-aware authority handling, and failure architecture for recovery. These capabilities span data, trust, authority, transaction, and recovery readiness.
AI-ready is an internal capability question - can your organisation use AI tools effectively? Agent-ready is an external interface question - can autonomous AI agents interact with your organisation effectively? Many organisations that are highly AI-ready are completely unprepared for external agent interaction because agent-readiness requires trust architecture, machine-readable content, and programmatic transaction capability.
Being agent-ready is not the same as being AI-ready. AI-ready typically means an organisation has adopted machine learning, automation, or generative AI tools internally. Agent-ready means the organisation is prepared for a world in which An agent-ready organisation has four characteristics: its offerings are These four characteristics map directly to the Many organisations that are highly AI-ready - using copilots, generative tools, and internal automation - are completely unprepared for the moment when their next customer arrives not as a human browsing a website but as an AI agent executing a delegated mandate. Agent-readiness requires a fundamentally different design posture: one built on Products, services, policies, and capabilities must be expressed in structured, machine-readable formats. Schema.org markup, JSON-LD, standardised product feeds, and clean API documentation are the foundation. If an agent cannot parse your offering, you do not exist in the agentic economy. Agents evaluate trustworthiness through verifiable signals - not brand stories. Uptime records, fulfilment accuracy, return rates, customer satisfaction scores, and compliance certifications must be queryable and auditable. Trust architecture replaces marketing as the primary mechanism for earning agent selection. Your systems must be able to receive, validate, and honour delegated authority from agents acting on behalf of humans. This means understanding delegation scope, mandate constraints, and the conditions under which an agent's authority should be accepted or challenged. End-to-end programmatic transaction capability - from discovery through evaluation, negotiation, purchase, and post-purchase - must be available via APIs and webhooks. Machine-to-machine checkout flows replace human-navigated funnels. When autonomous transactions fail - and they will - your systems must support machine-readable error handling, automated dispute resolution, and trust recovery mechanisms. Failure archite