This 3-hour hands-on course teaches UX designers and engineers how to move beyond traditional UI to build AI interfaces that make reasoning visible, steerable, and accountable. Learn the 5-layer trust architecture, apply actionable interaction patterns, and instrument metrics that turn user trust into a measurable product capability. Aligns with AEBOP T3.2.
AEI members save 20% with code MEM_C14_20.
This module deconstructs the core paradigm shift from task-oriented UX to cognition-oriented UX. You will learn why interfaces are no longer endpoints but become active "trust surfaces" where humans observe and steer autonomous reasoning. We establish the fundamental principle that in agentic systems, usability is redefined as oversight.
We introduce the key models and definitions that underpin the discipline, including the Human-Agent-System responsibility model and the cognition loop (Intent → Reason → Act → Reflect → Learn). You will leave with a clear, five-point checklist to evaluate whether a UX is truly agentic, moving from abstract theory to a concrete evaluation framework.
Here, we translate principles into structure. You will master the five-layer "Architecture of Trust" (Intent, Cognition, Control, Feedback, Framing) that replaces traditional UI stacks. Each layer is dissected to reveal its specific role in making cognition legible and governable, forming a closed trust circuit from goal capture to policy verification.
We then operationalize this architecture through a library of six proven design patterns, such as the Plan Preview and Reasoning Trace. You will learn how to map each pattern to specific trust principles and engineering hooks, transforming abstract layers into buildable, interactive contracts between user and agent.
Get hands-on with production-ready code. This module provides direct, adaptable implementations for core patterns, including WebSocket reasoning streams, REST control endpoints, and integrated feedback widgets. You'll set up telemetry from day one and learn to harden systems with idempotent controls, PII discipline, and CI checks for UX regression.
We then wire these components into the Cognitive UX Loop, the engine of continuous learning. You will design the bidirectional flow where user feedback directly shapes agent behavior and is visibly reflected back. The focus is on implementing a measurable, closed-loop system where correction transforms into capability.
Agentic UX must earn trust daily. We establish field-tested operational practices, prioritizing "visibility before aesthetics" and enforcing mandatory abort paths. You will deploy a readiness matrix to guide systems from prototype to production and define clear roles (Designer, Engineer, UXOps) for sustaining the trust loop live.
Finally, we define the lean metrics stack that matters: five core signals (Visibility, Explainability, Steerability, Feedback, Policy) that act as the heartbeat of UX health. By studying common anti-patterns and real-world case studies, you learn to diagnose trust erosion and cement the principle that UX is ultimately governance in motion.