Start Here: Agentic Experience Design
Explore Agentic Experience Design as the evolution of Conversation Design for AI systems with behavior, tools, memory, and decision-making. A start here page for conversational designers.
I created Agentic Experience Design as a section for people who want a clearer understanding of what changes when we stop designing only interfaces and start designing systems that can act, decide, retrieve, remember, and escalate.
A lot of the current conversation around AI starts with tools, frameworks, and model capabilities. But the real design questions begin somewhere else:
What kind of behavior should a system be allowed to have?
What should it do under uncertainty?
When should it act, ask, verify, stop, or hand over?
And how do we make that behavior trustworthy for users?
That is what this section is for.
This section is one of two editorial tracks in this publication.
Conversation Design Basics is for people moving into Conversation Design from UX Writing or content creation and looking for a clearer understanding of the discipline and its foundations.
Agentic Experience Design is the track focused on what happens next: the professional evolution of Conversation Design as systems gain behavior, tools, memory, and decision-making capacity.
The core idea behind this section is that the move toward Agentic UX should not be understood as a purely technical shift. It is a change of professional level for conversational designers. The work no longer consists only of designing dialogues, intents, or flows, but of shaping behavior, boundaries, decisions, repair, and trust in AI systems.
If Conversation Design helped us shape what a system says, Agentic Experience Design asks us to shape how a system behaves.
You can read the pieces in order, or start with the ones that feel most relevant to your work.
Start with:
What is Agentic UX Design? A Beginner’s Guide to Designing AI Agents
The End of the Flowchart: Why Conversation Designers are Feeling Lost
Brand Voice Isn’t a PDF. It’s the Rules That Control Your AI’s Decisions
From there, the section moves through questions of behavior design, trust, retrieval, escalation, memory, repair, and the changing role of conversational designers in AI systems that do more than respond.
Carmen, this distinction between 'what it says' and 'how it behaves' is the most important shift in design leadership right now.
For a long time, we’ve treated conversation design as a layer of polish—making the machine sound a little more human. But with Agentic UX, the 'design' is the logic of the system itself. If we’re only focused on the dialogue, we’re just auditing the output. The real craft is in designing the boundaries of uncertainty.
The hardest part of this shift isn't the technical complexity; it's the cultural one. In most product organizations, 'behavior' is still seen as an engineering problem, while designers are expected to stay in the UI. But if we don't own the rules for when a system should stop, verify, or escalate, we aren’t actually designing the experience—we’re just writing the captions for it.
Thanks for giving us a vocabulary for this next level of the craft.