
MDes Thesis Casestudy (2023-24)
My role
Product Designer, User Researcher- User research, Comparative Analysis, Persona Building, Journey Mapping, Storyboarding, User Flows, Wireframes & UI Design, User Testing
SUpervisor
Haig Armen
Timeline
4 Months
OVERVIEW
This innovative group trip planning application concept was an outcome of my MDes thesis project at Emily Carr University of Art & Design. I explored how AI can simplify group trip planning by understanding user needs and designing features to assist with coordination and collaboration.
The challenge
Highlights
Image
Relevant Itinerary Generation Flow
The algorithm analyses the preferences of the group based on the selected instagram posts and creates a relevant itinerary which serves as a starting point.
Image
Snap from the Group Chat
An AI chatbot that personalizes travel recommendations by learning your group’s preferences
Image
Snap from the Group Chat
Tracks participation and encourages input from less active members.
Image
Snap from the Group Chat
Zeelo detects conflicting signals based on the conversation, then offers assistance accordingly.
Image
Snap from the Group Chat
Summarises the discussion so far to align everyone's understanding. Suggests optimal approach to make a decision.
GIF
Ranking preferences drawer interaction
Group members rank their preferences in private
Image
Snap from the Group Chat
Zeelo analyses the groups responses and proposes options that align with the collective interests of the group.

Image
Two friends chatting across different timeszones

Image
A diagram representing the design process

Image
A diagram representing the design process

Image
Heuristic Evaluation
Prefer impulsive travel
Reach a consensus for all activities
Delegate decision-making to AI
Prefer flexible travel plans with some parts planned
Embrace separate plans according to the preference
Need help with facilitation but prefer to make the final decision
Slideshow
Journey Map and Personas
Research Summary
A well-rounded perspective that guided design decisions
Demand for Hyper Personalised Content
Ensuring relevancy in search recommendations
Asynchronous Collaboration
Enabling meaningful contributions without the pressure of real-time discussions.
Adaptability to Varied Schedules
Ensuring flexibility and adaptability to accommodate diverse schedules
Balanced & Effective Decision-Making Process
Ensuring equal recognition and participation for all voices
Inclusivity
Accommodating diverse preferences while preserving each other’s choices
Dynamic mental models
Reaching a mutually agreeable approach to travel planning

Image
Comparative Analysis

Image
Initial Sketch of Chatbot Facilitation in a Group Setting

Image
Iteration process for group chat-based AI facilitation
Initial Exploration
Users appreciated the concept of an AI chatbot assisting with navigating complex group discussions during travel planning. However, they found the presented design unclear and felt the name ‘Travisor’ lacked the catchiness they were looking for
Iteration 1
The app name ‘Zway’ and chatbot name ‘Zeelo’ were better received by users. However, they remained dissatisfied with the primary color and the low-fidelity prototype.
Iteration 2
The primary color was updated to purple, reflecting technology, mystery, and magic, while significant UI changes were made and highly praised by users
Iteration 3
Users raised an important question: wouldn’t it be simpler for them to arrange their preferences instead of accommodations, allowing the AI chatbot to analyze and provide tailored recommendations? leading to the final iteration.

Image
A Storyboard of AI chatbot facilitation through an example
1. Relevant and Personalized Recommendations
2. Analyse Group Needs, Problems & Conversation Dynamics
3. Promote Equal Participation
4. Facilitates Decision-making
5. Maintains Transparency
6. Ensures Privacy of Users

Image
Screens showing AI-chatbot facilitating group discussion for deciding accommodation
This project gave me a chance to think strategically about how to best place AI into a service and create a convincing prototype that helped to test the hypothesis.
User-centric approach to designing with AI
By gaining a deeper understanding of how AI works, I was able to think beyond just chatbots.
Continuous Iteration
User feedback offered qualitative insights into how the product was perceived and used. This iterative process helped identify pain points, optimize features, and ensure that the final design met the actual needs

Image
System Architecture
While not being able to interact with an actual AI-bot, participants understood the concept and responded positively
Participants were appreciative of AI's assistance in facilitating group chat and recommendation
While many were excited and optimistic about the prototype’s potential, yet some also expressed concerns regarding data privacy and security