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 |
Service design coordinates people, processes, touchpoints, and systems across a service journey. AXD becomes necessary when one of the participants is no longer a passive system but an autonomous actor with delegated authority. Service design assumes human actors navigating journeys; AXD designs for agents that hold goals, make choices, and act across time without constant human supervision.
No. AXD does not replace service design \u2014 it extends it into territory that service design was not built to cover. Service design remains essential for human-facing touchpoints, orchestration, and experience coordination. AXD adds the design of delegation, trust architecture, observability, and recovery for autonomous agents operating within those services.
Service designers need AXD when autonomous agents become participants in the service \u2014 not just tools that humans use, but actors that hold delegated authority and make decisions independently. This includes agentic commerce (agents shopping on behalf of customers), agentic banking (agents managing financial transactions), and any service where AI acts in the human's absence.
Autonomous agents change service design by introducing a new class of participant that operates without constant human supervision. This requires designing for delegation (how authority is granted), trust architecture (how confidence is built and maintained), observability (how humans monitor agent behaviour), and recovery (how the system responds when things go wrong). These are AXD design problems, not traditional service design problems.
Service Design, as a discipline, is fundamentally anthropocentric. It emerged from a need to choreograph the complex interactions between people and the services they consume, viewing the 'user' as a human actor with specific needs, goals, and emotional states. Its tools and methods, such as customer journey maps, service blueprints, and personas, are all predicated on the assumption that a human is at the center of the service experience, making decisions, performing actions, and experiencing touchpoints. This human-centricity has been its greatest strength, enabling the creation of more intuitive, efficient, and desirable services across countless industries. Service Design gives us a language and a process to see the invisible structures of a service, mapping out the frontstage and backstage interactions to ensure a seamless and coherent user experience. It excels at orchestrating the people, processes, and props that constitute a service encounter. The human is the 'principal' who delegates a task, but the agent is the one navigating the digital or physical environment to achieve the specified outcome. This distinction is critical. A user directly experiences a service's touchpoints, whereas a principal experiences the outcome of the agent's actions. Their experience is indirect, mediated by the agent's performance and reporting. This shift requires us to move beyond the concept of a 'user journey' and instead focus on the 'delegation architecture'. We are no longer designing for a person's interaction with a system, but for a system's interaction with other systems on behalf of a person. The agent's 'experience' is one of APIs, data protocols, and operational envelopes, not intuitive interfaces or emotional responses. Therefore, the design focus shifts from usability and satisfaction to Can the principal trust the agent to act in their best interest? Is the delegation's scope clearly defined? Are the constraints robust? How does the agent handle ambiguity or fa