OVERVIEW
An AI Chatbot Website that personalize honeymoon planning for newlyweds.
Advised by Camp Media, a startup focused on the honeymoon industry.
MY ROLE
I led end to end UIUX design including identifying user pain points, user journey analysis, design system and branding, wireframe and usability testing.
My prior experience with various AI chatbot platforms laid the groundwork for Honeybot's distinctive user flow design.
GOAL
Reduce honeymoon planning time.
Increase booking rate of affiliated resorts, activities and restaurants.
TEAM --- Cornell Tech
Mengying --- Designer
Sandhya --- Business Strategist
Amy ---- Engineer
Sherri ---- Engineer
Clara ---- Legal Researcher
DURATION
Aug - Dec 2023
MEDIUMS
Figma, Miro
Design Brief
How might we help newly-weds plan the perfect honeymoon?
Understanding the Problem
It takes too much time for couples to choose the destination and leverage information among multiple platforms without guidance.
Design Solution
Feature 1:
Tailored Recommendations and Guided Discovery
Honeybot leverages AI to deliver personalized honeymoon suggestions aligned with user preferences and budget constraints. To facilitate informed decision-making, users can explore curated recommendations through interactive guided experiences.
Challenge: Excessive time to perform research and lack of direction to where to start.
Design Solution
Feature 2:
Authentic Insights with Verified Reviews
Discover top-rated resorts, activities, and restaurants tailored to your preferences with HoneyBot curated packages, backed by AI-summarized reviews and direct links to firsthand experiences.
Challenge: Distrust towards travel agents and online advertising
Design Solution
Feature 3:
Effortless Itinerary Creation, Easier Bookings
Save favorite spots and generate personalized itineraries for each destination. Experience the simplicity of organization and seamless honeymoon booking.
Challenge: Burden on users to make lots of bookings.
Process
At the beginning, we created rich picture to explore, acknowledge and define the current honeymoon industry situation in order to analyze the relationship and potential partnerships between different stakeholders and players.
Stakeholder Analysis
1. RESEARCH - Understanding the Landscape
Competitive Company Research
We researched on how digital products in honeymoon industry evolve through time due to the development of technology and couples' behavioral change, from all-inclusive package into DIY.
Therefore, we envision our product to provide more personalized recommendations according to different needs, travel styles and financial background from couples.
Interviews
We interviewed with 12 couples at NYC City Clerk, 2 travel agents and 4 agents at the luxury resorts to understand the users needs.
“Honeymoon should be longer, more focused on the couple, there are sites, but more about having the time together, and more luxury.”
“Somewhere to see great reviews from other friends or similar-minded people, local networks.”
“We found an agent through word of mouth, a site recommended by a family friend specifically for the country we are traveling to.”
“It’s hard to know what to do on a honeymoon. So we look for itineraries from people that we know or from the internet”
Insights:
TRAVEL TRENDS
Couples prefer travel multi-destination & international
RECOMMENDATIONS
Couples want to see reviews and stories from other couples
TRAVEL AGENTS:
Couples looking for local gems from trusted advisors and agents
Pain Points
Journey Map
1.2RESEARCH - Understanding the Users
After analyzing the interviews, I created a journey map to spot pain points and emphasize user needs step by step.
I found that it would be overwhelming for couples to leverage and compare information among multiple platforms in different planning stages.
Persona
1.2RESEARCH - Understanding the Users
I created persona to narrow down our product customer segment.
Brianstorm
3. IDEATION
As a team, we came up with 100 ideas based on user pain points, market trend and technology innovation.
After many rounds of team discussion, expert consultation and de - risking analysis, we narrow down to one direction.
We created business model canvas for each venture to explore market fit and technically and economically feasibility.
And we narrowed down into 2 ideas. After de-risking the two ventures with the team and getting feedback from company advisor, we decided to move forward with Honeybot.
SWOT Analysis of Bard and ChatGPT
IDEA VALIDATION
In order to identify our differentiated features., we used Bard and ChatGPT to mimic the process of planning honeymoon and created a swot analysis of these chatbots, which are major competitors for honeymoon planning using AI
Insights
1.
The built-in capabilities from current LLM chats are not customized and specific enough for people planning honeymoons.
2.
The current chatbot answer UI design is not engaging and lack of photo assets.
3.
The needs to have a dashboard for trip collection, organization and chat history.
User Flow
3.1 IDEA VALIDATION
After analyzing insights from previous steps and discussion with team, I drew a user flow to map the user’s path within the product and communicate user journeys clearly to the team and stakeholders.
Low-Fi Protype
3.1 IDEA VALIDATION
I designed the low-fi prototypes to visualize product key features of AI conversation, AI answer card and dashboard for collection and itineraries.
3.1 IDEA VALIDATION
Wireframe
To conduct idea validation experiments to users, I drew the wireframe to visually represent the layout of each screen step within the user flow and set the grid system.
3.1 IDEA VALIDATION
Experiments
To ensure the product's viability, we identified 3 key potential risks.
To mitigate these risks, we designed targeted user experiments.
Utilizing wireframes and a pitch deck, we conducted these experiments with over 30 couples recruited through our internal network. A thorough analysis of the user feedback gleaned from these experiments informed subsequent design modifications implemented on the prototypes.
Design Iteration 1 ---- Addressing Diverse Planning Approaches
1. Implementation of guided prompts at the initial user interaction. These prompts would cater to the diverse needs of couples by allowing them to begin the planning process with the questions most relevant to their priorities.
2. Include structured questions prompting users to define their desired honeymoon experience at the outset. This modification aims to offer a personalized and efficient planning journey for all users.
(Before)
(After)
Design Iteration 2 ---- Prioritizing Reviews
We envisioned establishing a partnership with platforms like TripAdvisor to seamlessly guide users to relevant first - person review sections for in-depth information. However, acknowledging the potential for information overload, we determined to offer a two-pronged approach. The platform would still provide AI - generated summaries of reviews, allowing users to gain a general impression before delving deeper into specific user experiences. This dual approach empowers couples with the information they value most while ensuring an efficient planning experience.
(Before)
(After)
4. Design Solution
Visual Identity
Based on the user research and our product vision, I design the visual identity that provide calm and romantic feelings, also reflecting on the modern technology.
5. Conclusion
Next Step
Recent advancements in applications from industry players and technology(Open AI API & Gemini) signal that this product is a viable business idea. We want to integrate LLM models in out high-fi prototype to further test product’s risks of usage, where we could trace logs of participant activity, usage, click-throughs, and return rate to the application. Much of the user feedback relates to which sources of data populate the returned results, so it would be beneficial to test which sources or combination of sources for booking links and pricing are most trustworthy to users next.
