Conversation
to understand intent.
Use natural language when the customer’s goal is still fluid.
Domain Search & Registration Agent · 2025
From vague intent to a confident domain choice.
I designed an AI-guided experience that interprets an early business idea, recommends relevant domains, and helps customers move confidently toward registration.
- Role
- Product Designer
- Focus
- Agent UX · Interaction strategy
- System
- GoDaddy Airo
01
CONTEXT
AN ECOSYSTEM OF AGENTS
Airo knew the customer.
The Domain Agent knew the task.
GoDaddy was evolving Airo into a coordinated system of specialized agents. Airo maintained customer context and directed work to the right product expert.
My focus was defining how the Domain Agent should understand intent, present recommendations, and support a purchase—without becoming an interrogation or an endless chat.
JOURNEY ENTRY
The request begins in Airo.
Customers enter through the shared Airo home before the Domain Agent takes on the specialized task.
THE DESIGN CHALLENGE
“I want to buy a domain”
isn’t a complete brief.
The same request could come from a first-time founder, an established business, or someone adding a new venture. The agent needed enough context to be relevant—while minimizing questions and preserving control.
YOU
I want to buy a domain.
FIRST BUSINESSEXISTING VENTURENEW BRANDDOMAIN PORTFOLIO
02
STRATEGY
THE INTERACTION MODEL
Three modes for three kinds of thinking.
Conversation
Structure
to compare options.
Move names, extensions, availability, and price into a scannable surface.
Clear actions
to make a decision.
Reduce uncertainty as the customer moves from exploration to commitment.
DESIGN DECISION · 01
Reuse context before asking questions.
When Airo already knew about a customer’s ventures, the agent surfaced them as editable choices. Customers could confirm an existing business or start something new without repeating information.
Ask customers to verify known context—not recreate it.
DESIGN DECISION · 02
Gather intent progressively.
For a new venture, the agent asked only what would improve the next recommendation. Each answer became a visible signal—not information disappearing into an opaque model.
Every question should improve what comes next.
DESIGN DECISION · 03
Move comparison
out of the chat.
Conversation remained the place to express preferences. A dedicated results surface made domain names, extensions, availability, and pricing easier to evaluate.
One continuous task. Two coordinated surfaces.
A DECISION, NOT JUST A LIST
Move from exploration
toward commitment.
RESPONSIVE PRODUCT DESIGN
The decision model
held up on mobile.
The same hierarchy carried into narrower surfaces: the recommendation remained scannable, the value proposition stayed close to the domain, and the next action remained within reach.
THE PROTOTYPE
The interaction,
in motion.
The full prototype brings the agent’s conversation, recommendation surface, and refinement loop together in one continuous journey.
DESIGNING THE WHOLE EXPERIENCE
Design for the moments
between the ideal moments.
Waiting, refinement, recovery, and handoff were treated as part of the core journey—not edge cases. When the system could not respond, the experience explained what happened and kept retry close at hand.
- Empty and onboarding
- Search and refinement
- Error recovery
- Registration handoff
OUTCOME
A repeatable model for
agent-led decisions.
The work established a clear pattern for AI-assisted commerce: use conversation while intent is fluid, introduce structure when comparison begins, and preserve context through every transition.
Designing the agent was less about making chat feel intelligent—and more about deciding when the interface should listen, guide, reveal structure, or get out of the way.