EVA is an AI powered concierge designed for eCommerce websites, where the primary interface is conversational rather than visual. The experience focuses on voice and chat interactions, supported by a minimal UI consisting of an agent presence, audio input, and a chat window. My goal was to design an intuitive, human like interaction model that guides users through shopping decisions with clarity and confidence.
When interfaces disappear, interaction design becomes the experience.
Traditional eCommerce experiences rely heavily on complex navigation, filters, and dense product listings. When introducing an AI concierge with a minimal UI, the challenge shifts from visual design to interaction clarity how users initiate conversations, understand system responses, and feel guided without visual overload. A poorly designed conversational flow could easily confuse users or feel robotic.
Understanding how users seek help when there’s nothing to click.
Research showed that users expect AI assistants to be conversational, contextual, and reassuring especially when the interface is minimal. They prefer clear prompts, visible listening feedback, and responses that feel proactive rather than reactive, reinforcing the importance of trust and clarity over visual complexity.
Designing conversations as the primary interface.
I explored voice-first and chat first interaction patterns, mapped key eCommerce moments where users hesitate, designed conversational prompts that reduce decision fatigue, and focused on making system feedback visible through subtle cues like listening states, response timing, and tone of voice.
Here comes EVA…a calm, human like concierge where conversation
becomes the interface.
I designed EVA around a minimal,
distraction free interface anchored by an agent presence, voice
input, and a conversational chat window. Clear listening indicators,
natural response flows, and contextual guidance help users feel
understood and supported, turning complex shopping decisions into
simple, guided conversations.