Product designer • Project duration 1,5 months
Disclaimer
To comply with non-disclosure agreements, the following case study focuses on design processes, methodologies, and strategic outcomes. Sensitive information, client name, and proprietary data have been generalized.
Client overview
Fly Beyond is a travel & lifestyle app for a major MENA airline, combining an AI concierge, a virtual card and loyalty miles.
Problem
The airline wanted a concierge that could actually plan and book in conversation, not a chatbot bolted onto the app. It also had to fit together with a virtual card, two types of miles, a marketplace, trips and events without overloading the navigation.
My role
Product designer, responsible for the informational architecture of the app, a conversational AI concierge experience and a few flows, related to the virtual card, a clickable prototype for the client.
Key decisions
Making the concierge feel like a concierge, not a chatbot
Challenge: The core job was "handle everything for me, so it feels effortless and premium." A text chat asks the user to type requests and read long answers, often while they're on the move during a trip.
Decision: I proposed voice as the primary mode, with text as a fallback. In a trip context, one voice conversation plans a full day, books it, adds it to the calendar and suggests the next step, like a taxi or a spa.
One concierge, two very different contexts
Challenge: At home, users want "something nearby tonight." Before or during a trip, they want a plan for an unfamiliar city. Treating both the same meant users had to explain their context every time.
Decision: I split the concierge into home and travel modes. It started as a manual toggle and evolved into automatic switching when a flight out of the user's city appears in their calendar.
Two types of miles, one tier
Challenge: Miles (from flights) and Points (from marketplace purchases) work differently but add up to one shared tier. The accrual logic lived in another of the airline's apps, and showing it as-is would confuse users.
Decision: I mapped the existing logic in FigJam and aligned it with the team, then designed a tier screen that shows one total with a clear breakdown by source.
Five product areas, one bottom bar
Challenge: The concierge, card, marketplace, trips and events all competed for navigation, and the visual concept of the home screen was already fixed.
Decision: I explored two informational architecture concepts: concierge-first with a side menu, and a bottom bar with a feed. The recommended version merged them: a central concierge button in the bottom bar, a calendar instead of trips, and the marketplace as its own tab.
One prototype that holds together as a real product
Challenge: The client asked us to combine everything into a single prototype to present to their own customer. The flows had been designed separately, so the screens didn't share one story. The same user, trip, balance and bookings had to stay consistent from screen to screen, or the demo would fall apart.
Decision: I built one end-to-end scenario, from pre-onboarding through home, card, concierge and trip to marketplace, and aligned the content across it, so a booking made in the concierge showed up in the calendar and the miles matched everywhere. I used smart animate and delays so the transitions felt like a real app, not clicking through static screens.
Impact
The client presented the prototype to their end customer and got positive feedback. Seeing it end-to-end let both of them read the product as one system rather than a set of screens.