— The Boston Globe
Overview
Mar 7 - May 2, 2024 (8 weeks)
Timeline
Designer, Researcher, Product Manager
Role
Arkadiy Avrorov, Logan Perez
Advisors
Figma, Google Forms, Zoom, Asana
Tools
Background
This project was inspired by my personal experience arranging an international trip for my parents to visit me in San Francisco. They were thrilled at the thought of family reunion, but I soon noticed how quickly excitement turned into stress.
Flights, visa paperwork, and countless small decisions piled up, leaving them more anxious than joyful. That experience made me wonder: how might we make travel planning feel less like a burden and more like part of the adventure?
Responsibilities
This is an independent project for my Capstone project at Maryland Institute College of Art. As the sole designer, my key responsibilities include:
Conducted user research to develop deep understanding of the user's needs, behaviors, and pain points.
Conducted market research to identify service gaps and opportunities.
Created wireframes and high-fidelity prototypes to visualize the product's concept, layout, and user flow.
Conducted usability test to identify usability issues and areas for improvement.
Collaborated with engineering consultants to validate technical feasibility.
Own and manage the project roadmap to ensure on-time delivery.
Project Vision
Through this project, I want to make travel more accessible for individuals of all ages and levels of digital literacy. My vision is to empower people to explore the world with confidence, and to reduce the friction and uncertainty that too often overshadow the joy of planning a meaningful journey.
The Problem
Travelers often feel stressed over trip planning tasks due to information overload. From countless options for flights, hotels and activities to endless travel content on social media, the process of making travel decisions has become increasingly complex.
Discover the Problem Space
I adopted a comprehensive research strategy to unfold this issue from both the user and business perspectives. Through online surveys and user interviews, I gained firsthand insights into the travelers' preferences, expectations, and frustrations. Through market research, I examined existing solutions, pinpointed their strengths, weaknesses, and opportunities for TravoAI to innovate. This dual approach allowed me to gather actionable insights and shape TravoAI to not only meet the needs of my users, but also differentiate itself in today's crowded tech market.
From the user's perspective
Survey
Survey results are based on 118 qualified responses from participants over 18 years old, who has planned at least one trip during the past 12 months.
Traveler type
45% are vacationers
28% are digital nomads
26% are backpackers
Travel frequency
96% travel 2-4 times per year
2% travel once per year
2% travel 5+ times per year
Travel group size
50% travel in small groups (2-5)
28% travel alone
22% travel in large groups (6+)
Source of travel contents
Online travel agencies (OTAs)
Search engines
Social media
Stressful tasks
Securing flights /lodging
Deciding location transportation
Managing budget
Primary tools
Smartphone
Desktop/laptop
Tablet
User Interview
To enrich the data gathered from my online survey, I also interviewed 5 participants from 28 to 66 years old to gather qualitative insights on their trip planning behaviors and the challenges they face throughout the journey.
I've come to dread the routine of the family vacation. It's not that I don't want to spend the money or certainly spend time with my wife and kids... but the annual stress involved with trying to figure out coordinating planes, when we have to be where, what places we want to visit, and then getting there and having to deal with the unknowns of where to park, what the local requirements are, etc. It's all just too much.
— Jonathan, 57
Planning trips stress me out so much because there are too many variables.
— Emily, 42
I probably wouldn't enjoy trips that need a lot of planning.
— Sarah, 61
Uncover user pain points
Based on our conversations, several key themes came to light regarding the challenges participants faced while planning their trips.
Starting point
Not every traveler can answer the "Where to?" question
Most participants begin planning their trips without a specific destination in mind, relying heavily on research and inspiration to guide their decisions. However, many major Online Travel Agencies (OTAs) and itinerary planning tools require a predetermined destination to initiate the planning process. As a result, users often turn to third-party resources to explore travel content.
Stress trigger
Travel booking is the most "painful" stage
In the early stages of planning, research is more spread out and relaxed. However, as participants approach the date of booking, they experience a significant increase in both travel content consumption as well as their stress level due to the amount of decision-making involved.
Duration
Scout the internet for hours to build the perfect itinerary
Most participants start trip research 6 to 12 months in advance to secure the best deals on flights, lodging, and activities. On average, they spend over 10 hours on travel content prior to booking a trip and dedicate an additional 2 to 5 hours to creating a detailed itinerary.
Define user values
Conducting both qualitative and quantitative research allowed me to capture a wide spectrum of insights on the things users value the most when they interact with travel planning products.
Efficiency
Users are overwhelmed by managing travel content across platforms and appreciate a centralized tool that streamlines destination recommendations, route planning, booking, budgeting and more.
Personalization
The diversity in travel styles highlights the importance of integrating personalized travel content, particularly allowing flexible output and granular customization to meet specific user needs.
User-generated content
There is a clear desire for travel contents considered trustworthy, whether through personal connections or social media platforms with rich user-generated contents.
Collaboration
Majority of users travel in groups, therefore it is important to allow collaboration between travel group members.
Ease of use
An intuitive interface is crucial to ensure a positive experience for users across all ages and digital competencies.
Personas and journey maps
By integrating personal narratives and statistical trends, I developed two user personas that capture the distinct travel planning styles and motivations as identified in my research. I positioned these personas within their respective journey maps to provide a holistic view of their experiences, key touchpoints and feelings at each phase of travel planning. This method allowed me to identify the dull moments of their user journeys and discover opportunities for improvement.
Meet Eric
An adventurous young professional looking for travel ideas to inspire his next vacation. Eric is motivated to travel because it is a channel for him to relax and build new experiences with family.
Meet Mary
A retired traveler who needs to plan a trip to a known destination. Her motivation to travel has changed over age and she finds joy in returning to familiar places to retrieve old memories.
From the market's perspective
Competitive Analysis
I centered my market research around evaluating key competitors' ability to meet user values identified through user research. This method allowed me to know my competitors' strengths and weaknesses, and identify service gaps to foster innovative problem-solving.
Conclusion
AI enhances experiences, not replacing them
AI alone does not solve the problem. While it can process vast amounts of data and provide recommendations quickly, AI lacks the human touch necessary to adapt to the subtle preferences and emotional needs of individual travelers. Effective solutions require a blend of AI-driven efficiency and human-centric design, ensuring that the technology serves to enhance, rather than replace, the user experience.
AI's effectiveness in trip planning is also limited by its lack of access to real-time data. For travel planning to be truly responsive and accurate, it needs up-to-the-minute data on factors like airline ticket prices. Without this information, AI can only provide recommendations based on outdated or static data, potentially leading to less reliable solutions. Integrating real-time data feeds into AI systems is essential to enhance their utility and ensure they meet the dynamic needs of our travelers.
How Might We
HMW help travelers to easily navigate information overload during travel research and minimize their cognitive effort when making travel purchases?
Break Down the Problem
The ideation phase was a crucial step in transforming initial concepts into actionable solutions. Starting with a mind map,
rapid ideation, I further broke down the HWM and generated a diverse range of possible ideas to address the core challenges identified in my research. To systematically evaluate and prioritize these ideas, I employed the Kano Model so I can focus on the most impactful solutions when designing the minimum viable product.
Rapid Ideation
How might we help travelers to automate travel research?
How might we empower travelers and make them feel more confident about traveling?
How might we make collaboration easy between travelers?
Prioritization
To avoid biased design decisions and ensure that TravoAI prioritizes feature that aligns with user needs, I implemented a Kano analysis during the ideation phase. The following analysis was based on feedback from 5 participants.
By identifying Basic, Performance, and Delighter features, I was able to strategically focus on those that not only meet basic user expectations but also provide an exceptional experience to differentiate our MVP from its competitors.
User Flows
Mapping out key user flows allows me to visualize the paths a user takes to accomplish their trip planning tasks.
Wireframes
I created low-fidelity wireframes to establish the skeletal structure of the app. These wireframes allowed me to visualize key user flows, explore different layout possibilities, and validate functionality before moving into high-fidelity design.
The Solution
TravoAI is a progressive web app designed to automate travel ideation and streamline trip planning tasks through a conversational interface.
Core experience
Create account
Incorporate user preference in sign-up
Although an account is not required to use TravoAI, I want to offer interested users a more personalized travel planning experience as well as the option to save their work for future access. This simple 3-step approach not only reduces friction during sign-up but also paves the way for a tailored trip planning journey.
Core experience
Itinerary Generator
Automate itinerary building with conversational AI
Users can engage with the trip planning assistant through a conversational interface to achieve specific travel planning needs. In addition to text and voice input options, the immersive voice mode simulates the feel of speaking to a travel agent in real time. This design significantly enhances communication efficiency and the overall itinerary building experience .
Add human touch to AI-generated content
The automation and continuous learning capabilities of AI systems allow users to scale processes efficiently and easily. However, AI is limited in its ability to understand context and predict specific user actions, requiring human intervention for optimal performance. Therefore, the goal of AI should be to enhance experiences, not replace them. With TravoAI, users can further customize AI-generated content to match their preferences and invite collaborators (co-travelers) to create their dream trip together.
Core experience
Travel Booking
Recommendations based on user preference and trip itinerary
Once travelers have narrowed down their trip options, the next challenge they face is booking accommodations, flights, and activities. TravoAI streamlines this process by analyzing user preferences and travel plans to suggest the most suitable options that align with their itinerary and budget, allowing travelers to shop their upcoming journey with minimal friction. Additionally, users can import reservations booked through other platforms, using TravoAI as a centralized hub for storing and managing all their travel booking confirmations.
Core experience
Explore Trips
Get travel inspiration from the community
User research revealed that most participants begin planning their vacations without a specific destination in mind, relying heavily on social media and word-of-mouth recommendations to guide their decisions. To address this need, explore trips feature allows travelers to discover trending destinations and easily adopt another user's itinerary to reduce research effort.
Iteration
User Testing
To validate the usability of my prototypes, I conducted 2 rounds of usability test with 5 participants (ages 28–67) that represented both tech-savvy travelers and seniors with lower digital literacy. Each participant was asked to complete the following core tasks:
Generate a travel itinerary and make customizations
Share itinerary with a travel mate
Find a community trip and turn it into their own
Design Revision
Issue 1
Issue 1
Function Discoverability
Issue 1
Function Discoverability
Finding
3 out of 5 participants struggled to find the share and customize itinerary features.
Why is that?
Customization and share features are hidden under the meatball menu, making them hard to discover.
The icon lacks a clear visual cue that it contains personalization tools.
Key actions like edit and share are treated as secondary rather than primary actions.
Users expected to see key actions inline with the itinerary cards.
Improvements
Results
In the follow-up test, 5/5 participants were able to locate and use the customization and share functions without guidance.
Task completion time for sharing itinerary reduced by 33%.
Task completion time for customizing itinerary reduced by 58%.
Participants described the new flow as “easy to spot” and “much faster.”
Issue 2
Copy clarity
Issue 1
Function Discoverability
Finding
All participants were able to copy a community trip, but 2/5 participant expressed confusion about the "Steal itinerary" button label.
Why is that?
The word “Steal” created a playful but negative tone that didn’t clearly communicate the intended action.
Participants were unsure if they were simply saving someone else’s trip versus making a copy for themselves.
The lack of supporting context (tooltip or helper text) increased ambiguity.
Improvements
Success metrics
In the follow-up test, all 5 participants immediately understood the purpose of the button.
Task completion time reduced by 12%.
Confidence in using the community trip adoption feature increased by 28% (self-reported).
Participants described the new label as “clear,” and “straightforward.”
Reflection
Lessons
Use straightforward, action-oriented language
Through user testing, I learned how critical it is to make key functions immediately visible and to use straightforward, action-oriented language. When features were hidden or copy was ambiguous, participants hesitated and lost confidence in the product. By surfacing customization and share options and clarifying button labels, I was able to reduce friction and help users feel more in control of their trip-planning experience.
Thoughtful survey answers are critical for good data
From synthesizing online survey results, I've learned that thoughtfully crafted answers are critical for gathering high-quality data. Specifically, when survey responses cover too broad a range, they can significantly dilute the specificity and usefulness of the data collected. This lack of granularity can obscure vital insights into the respondents' true experiences and opinions. To avoid this, it's crucial to design survey answers that are specific enough to capture detailed and actionable data, ensuring that each response can effectively contribute to a precise analysis.
Challenges
Prototype voice interaction in Figma
TravoAI leverages conversational artificial intelligence to understand and respond to human speech, enabling it to interact with users in a natural and intuitive way. Given that Figma does not support voice interactions, I utilized text-to-voice technology combined with video editing technics to create a realistic demo of the itinerary generation flow. This approach allowed me to simulate the voice interaction feature and showcase the product’s capabilities in a dynamic and engaging way.
Build custom ChatGPT takes time and expertise in ML
In this project, I adopted GPT4.0 to design the conversation and travel itineraries in my user cases. If time is permitted, I would like to explore custom versions of ChatGPT that combine instructions, extra knowledge, and combination of skills that is specific to trip planning.
Next Steps
01
Improve budgeting feature
Based on user feedback, TravoAI's current budgeting tool effectively tracks spending in three specific categories: flights, lodging, and tickets for activities. However, users have expressed a need for a more comprehensive financial tracking system that covers all aspects of their travel expenses. Moving forward, my focus will be on refining the budgeting tool to enable users to track or log all of their travel-related expenses.
02
Create custom GPT for travel planning
AI language models like ChatGPT are proficient at creating geographically and thematically cohesive travel plans, but how might we make the human-AI interaction feel more natural and less robotic? To address this challenge, I believe training a custom GPT for TravoAI is essential to enhance conversational capabilities and deliver a more human-like travel planning experience for it's potential users.
