Six disciplines that make agentic transformation real. The first is not "the agent." The last is not "the prompt." Between them sits the work that decides whether your AI investment produces P&L or a graveyard of demos.
Read the founding statement →We were here before the term “UX” was common. We’re here now, as AI makes human-centered thinking the most consequential skill in technology. The future is being designed.
The agentic era doesn’t replace design—it clarifies what designers are for. The Agentic Design Stack organizes six disciplines you already practice—service design, information architecture, interaction design, conversation design, operational design, and organizational design—into a map for where your judgment matters when agents enter the workflow. It’s how practitioners find their footing: not by chasing every new model, but by shaping the services, interfaces, and organizations that make AI accountable to people.
Each answers a different failure mode when AI lands in the enterprise. Open a discipline for a short take; category pages on From the Practice collect related articles.
Without service design, agent programs ship slices of workflow instead of a redesigned service. Each agent demos well; nothing adds up to P&L because nobody owned the end-to-end promise—the citizen, claimant, or patient experience—with humans and machines arranged inside one coherent service.
Information architecture structures knowledge, context, and operational state so agents compose instead of sprawl. It defines what a case, account, or consent is under what conditions—so incremental rollout compounds instead of producing fourteen islands with fourteen context models.
Human-in-the-loop queues fail when approvers cannot trust or use them. Interaction design shapes the handoff—what is summarized, what evidence ships with it, what the person is attesting to—so people actually run the loop and agent authority can grow legitimately over time.
Conversation design is not clever prompts; it is how dialogue is structured across channels so employees and customers stay in the flow. When people route around the agent, operational metrics quietly collapse—no matter what the dashboard claims.
Operational design (with systems thinking) decides how orchestration, composition primitives, and the agent runtime hang together—whether you can start small and grow without re-platforming every workflow. Without it, pilot two tends to rip out pilot one.
Agents change roles, accountability, and power. Organizational design names who manages an agent, who owns the decisions it encodes, and how teams evolve—before the program meets the union, the line manager, and risk. Without that theory, the other disciplines hit a workforce with nowhere to land.
Invisible Machines is the editorial engine of UX Magazine. Every episode is a conversation with someone at the frontier—researchers, executives, economists, neuroscientists, designers—exploring what it means to build intelligently in the AI era.
After each recording, Josh and Robb identify the core concept the conversation surfaced. That concept becomes an Ideation: the intellectual anchor for a cluster of related content—a guest article, curated further reading, and a members-only essay that synthesizes threads across multiple episodes.
This model exists because the ideas that matter most are rarely the ones that fit into a news cycle. They emerge slowly, in conversation, and reward the kind of sustained attention that a podcast enables and a structured archive preserves.
Robb has spent his career at the intersection of experience design and technology. As founder of one of the first experience design agencies (Effective UI), he built a practice that rivaled IDEO and Frog Design—then sold it to go deeper into the technology side of the problem.
He founded OneReach.ai—Gartner's top-rated platform for orchestrating conversational AI—on the conviction that the most powerful technology is the kind that gets out of the way. UX Magazine is part of the same mission: ensuring that as AI becomes infrastructure, the people building it remain accountable to the humans using it.
Josh is the editorial voice of UX Magazine and co-author of Age of Invisible Machines. His writing has appeared in The New York Times, Harvard Business Review, Raconteur, and Thrasher—a range that reflects a conviction that good ideas belong in every room, not just the ones where people already agree.
On Invisible Machines, Josh runs the post-episode breakdown where each conversation is distilled into its core Ideation. He is the person translating dense technical and strategic terrain into ideas that practitioners can apply—and the editorial conscience keeping the publication honest about what AI actually means for people.
UX Magazine has always been a collaborative publication. We have no corporate owner, no paywall on our essential content, and no single editorial voice. What we have is a global community of leaders, thinkers, and practitioners who share the belief that experience design has the power to make technology more accessible, more useful, and more humane—for everyone on the planet.
Join the community →UX Magazine has been built by practitioners, for practitioners. If you are working at the intersection of experience design and AI—or anywhere in the broader territory of how humans and technology meet—we want to hear from you. Guest articles, book excerpts, research findings, and field reports are all welcome.
Contributor guidelines →