An AI voice agent for Nissan Dubai that answers, qualifies and books, and turns every call into data.
Context
Nissan Dubai took sales and support calls all day, and every call was a lead, but none were being read. Managers could not see what customers asked, where follow-ups were lost or which test drives were booked.
Market Caller is an AI voice agent that answers those calls, qualifies the caller, books test drives and turns every call into data. I led the product design end to end: the call-to-insight journey, the interface, the brand system and the handoff to engineering, working with product, engineering and the client.
Decisions
I started from real calls: what customers ask and where follow-ups were lost. Then I mapped one flow with four stops: call, chats, analysis and overview.
Four decisions shaped it. Make the AI visible, with a live waveform and timer. Turn talk into clear fields instead of transcripts. Keep the interface calm, with signal colour only where it matters. And let a manager go from the overview down to a single call in a few clicks.
Design
Every screen from the live call to the dashboard: navigation, sign in, calls, chats, analysis fields, totals, funnel and settings, on one visual system.
Ink and violet lead, green only means good news and coral flags what needs attention. Plus Jakarta Sans and Funnel Display keep it human.
Takeaways
Every call tracked and understood: inbound and outbound in one place, missed calls flagged with an owner, and test drives booked from conversation to calendar.
Designing early for AI taught me that trust comes from showing the system's work. A manager believes a summary when they can open the call behind it.
