Case Study
Ask Abu Dhabi
Ask Abu Dhabi
Overview
Creating an AI-Assistant for Experience Abu Dhabi
To integrate AI into the Visit Abu Dhabi ecosystem, we designed a conversational assistant that helps people plan what to do—fast. My remit: define the end-to-end product thinking and state design for the chatbot across desktop and mobile, from first keystroke to saved itineraries.
Client
Experience Abu Dhabi
Year
2025
Services
UX/UI Design • User Journeys
Overview
Creating an AI-Assistant for Experience Abu Dhabi
To integrate AI into the Visit Abu Dhabi ecosystem, we designed a conversational assistant that helps people plan what to do—fast. My remit: define the end-to-end product thinking and state design for the chatbot across desktop and mobile, from first keystroke to saved itineraries.
Client
Experience Abu Dhabi
Year
2025
Services
UX/UI Design • User Journeys
Section
Meticulous by Design
Early reviews made it clear this needed to be more than a chat bubble. I built a living spec in Notion—every feature, every state, every edge case—so nothing was left to chance. That included:
Input states: idle, focus, typing, tags/filters applied
System states: retrieving, composing, confidence low, follow-up prompts
Result states: suggestions, rich cards (places/events), deep links, “add to favorites/itinerary”
Safeguards: no-results patterns, clarifying questions, graceful error handling
Each state was prototyped so stakeholders could feel the rhythm of the assistant: when it nudges, when it waits, when it shows options, and when it hands the next step back to the user.
Twelve Lives, One Brain
The client also asked for user stories spanning 12 personas—from luxury shoppers to cruise visitors and event-goers. I mapped which features mattered most to each, then wrote task-based journeys:
Cruise visitor → map-led discovery and short-stay bundles
Luxury traveler → rapid “add to itinerary” for multi-stop shopping and dining
Event-goer → “what’s on next” + wayfinding and transport prompts
First-timer / family → safe options, time windows, and save-for-later flows
This matrix kept the assistant context-aware without bloating the interface.
Section
Meticulous by Design
Early reviews made it clear this needed to be more than a chat bubble. I built a living spec in Notion—every feature, every state, every edge case—so nothing was left to chance. That included:
Input states: idle, focus, typing, tags/filters applied
System states: retrieving, composing, confidence low, follow-up prompts
Result states: suggestions, rich cards (places/events), deep links, “add to favorites/itinerary”
Safeguards: no-results patterns, clarifying questions, graceful error handling
Each state was prototyped so stakeholders could feel the rhythm of the assistant: when it nudges, when it waits, when it shows options, and when it hands the next step back to the user.
Twelve Lives, One Brain
The client also asked for user stories spanning 12 personas—from luxury shoppers to cruise visitors and event-goers. I mapped which features mattered most to each, then wrote task-based journeys:
Cruise visitor → map-led discovery and short-stay bundles
Luxury traveler → rapid “add to itinerary” for multi-stop shopping and dining
Event-goer → “what’s on next” + wayfinding and transport prompts
First-timer / family → safe options, time windows, and save-for-later flows
This matrix kept the assistant context-aware without bloating the interface.
Section
Designing the Conversation, Not Just Screens
Beyond UI, I treated conversation like a system:
Intent coverage: activities, visas, taxis/transport, opening hours, events, neighborhoods
Response choreography: concise first answer → scannable cards → one-tap next step
Memory & continuity: recent searches, soft personalization, and quick re-entry points
Multi-device parity: inputs and outputs tuned for both mobile ergonomics and desktop depth
Results That Speak for Themselves
Post-launch, the revamped website helped LearnPoint achieve significant growth in engagement, while feedback highlighted how both students and educators found the platform more intuitive and effective.
Section
Designing the Conversation, Not Just Screens
Beyond UI, I treated conversation like a system:
Intent coverage: activities, visas, taxis/transport, opening hours, events, neighborhoods
Response choreography: concise first answer → scannable cards → one-tap next step
Memory & continuity: recent searches, soft personalization, and quick re-entry points
Multi-device parity: inputs and outputs tuned for both mobile ergonomics and desktop depth
Results That Speak for Themselves
Post-launch, the revamped website helped LearnPoint achieve significant growth in engagement, while feedback highlighted how both students and educators found the platform more intuitive and effective.
Interested in working together? Call me.
Interested in working together? Call me.
Interested in working together? Call me.