UX
Architecture
About Me
UX
Architecture
About Me

AI Travel Agent

Personal project, 2023

Redesigned Google’s travel page to make travel planning effortless and effective.

Thank you for your interest, for better viewing experience, please consider opening in desktop.

Duration

12 weeks

Platform

Desktop

My Role

User research

UX Design

Prototype

Usability Test

Project

Background

Project Goal

This project started with me as a frustrated user. A week before my vacation to Mexico City, I opened ChatGPT hoping to have the AI genie save me some time by creating an itinerary for me. In a few clicks, ChatGPT generated a seemingly detailed itinerary. I saw the lengthy text and thought: oh well, what can go wrong? And then closed my browser and went back to my busy schedule.

Fast forward to a week later, I landed in Mexico, only to find out the AI-itinerary had me darting around a huge city with little logic nor guidance, and led me to generic over-touristic places that held no appeal to me.

The experience provoked me to start this personal project: How can we enable users to harness the vast potential of AI in generating itineraries that are not only efficient but also tailored to their preferences?

Problem

Statement

Problem Statement

How can we design the user experience to evoke confidence in AI-generated itineraries, ensuring they are intuitive and user-friendly?

Target

Audience

Target Audience

People who travel, age 14 to 65, who want a streamlined planning process that cuts down research time, as well as trustworthy and personalised recommendations.

Project

Outcome

Project Outcome

8 out of the 9 usability test participants expressed interest in using the product for their next trip.

Invited to present case study at Kiwi.com's Global Travel Jam hackathon

Design Approach - Design Thinking

Design Approach - Design Thinking

Research & Empathize

Market Research

Competitive Audit

User Interview

Understanding the market

Upon receiving the prompt, I conducted research to o understand the Post-pandemic tourism market.

Understanding the market

Upon receiving the prompt, I conducted research to o understand the Post-pandemic tourism market.

Quantitative Research

Quantitative Research

63%

63%

63%

Americans plan to take a leisure trip in 2023 [1]

Americans plan to take a leisure trip in 2023 [1]

$3000

$3000

$3000

Average budget for marquee summer trip [2]

Average budget for marquee summer trip [2]

3.1 Trips

3.1 Trips

3.1 Trips

Average planned trips per household for 2023 [2]

Average planned trips per household for 2023 [2]

5 Weeks

5 Weeks

5 Weeks

Before departure travellers are searching their destination for the first time [3]

Before departure travellers are searching their destination for the first time [3]

16x

16x

16x

Average number of times a user checked the price of their trip before booking on the Hopper app. [3]


Average number of times a user checked the price of their trip before booking on the Hopper app. [3]

68%

68%

68%

participants said that taking time to research before booking is more important now when planning their trips for 2022 or beyond than before the COVID-19 pandemic. [4]



participants said that taking time to research before booking is more important now when planning their trips for 2022 or beyond than before the COVID-19 pandemic. [4]

participants said that taking time to research before booking is more important now when planning their trips for 2022 or beyond than before the COVID-19 pandemic. [4]



Market Trend

Last-minute Booking

Last-minute Booking

Today, travellers are starting the travel planning process nearly 1.5 weeks closer to their departure dates for domestic trips, about a 30% reduction in travel planning time compared to pre-pandemic. [3]

Today, travellers are starting the travel planning process nearly 1.5 weeks closer to their departure dates for domestic trips, about a 30% reduction in travel planning time compared to pre-pandemic. [3]

No-normal Destinations

No-normal Destinations

Travellers are seeking unique destinations and experiences despite slowed economy. [2]

Travellers are seeking unique destinations and experiences despite slowed economy. [2]

Freedom to Roam

Freedom to Roam

20% Traveler say that they intend to work at least partially on their longest summer trip. [2]

20% Traveler say that they intend to work at least partially on their longest summer trip. [2]

Sustainability’s Imperative

Sustainability’s Imperative

62% of corporate travel managers say they expect their companies’ green initiatives to result in a reduction in travel of more than 25% by 2025. [2]

62% of corporate travel managers say they expect their companies’ green initiatives to result in a reduction in travel of more than 25% by 2025. [2]

American Travel Association, 2023 Travel Outlook

Deloitte, 2023 Travel Industry Outlook

Hopper, 2023 Travel Trends Report, Post-Pandemic State of Travel & Consumer Spending

Trip Advisor, 2023 Travel Trends Report

User Research

I held in-depth interviews with 5 people to understand what constitutes a good travel planner, and what is not working

User Research

I held in-depth interviews with 5 people to understand what constitutes a good travel planner, and what is not working

Questions:


  • How would you typically plan for a trip?

  • What are the apps/websites that help you plan for your trip? Why?

  • How do you and your friends plan a trip together?

  • To you, what does a good itinerary consist of?

  • How would you use your planned itinerary?

  • What are the biggest obstacles when you are planning for a trip?

  • What do you like the most about planning for a trip?

  • I asked each participant to use Chatgpt to generate an itinerary for their own city, and asked them about their likes and dislikes

Interview Insights

Themed Insights(Affinity Map)

Pre-interview biases and assumptions:


With new AI technology such as ChatGPT, travel planning should be easier than ever.


What’s verified:


People are interested in trying the AI-generated itineraries, 3 out of 5 interviewees have used ChatGPT to generate their itinerary, the others are planning to use it for their next trip


What’s refuted:


There isn’t enough trust on the current AI algorithm to produce personalised itinerary. The convenience of AI-generated itinerary can be a short-term relief to people who do not have time to research, however many people find it unreliable, and prefer to plan the old school way.

Pre-interview biases and assumptions:


With new AI technology such as ChatGPT, travel planning should be easier than ever.


What’s verified:


People are interested in trying the AI-generated itineraries, 3 out of 5 interviewees have used ChatGPT to generate their itinerary, the others are planning to use it for their next trip


What’s refuted:


There isn’t enough trust on the current AI algorithm to produce personalised itinerary. The convenience of AI-generated itinerary can be a short-term relief to people who do not have time to research, however many people find it unreliable, and prefer to plan the old school way.

Pain Points:

Pain Points:

01/


AI-generated itineraries are often unreliable, and lacks practicality. Users end up having to recreate their itinerary during their trip.

01/


AI-generated itineraries are often unreliable, and lacks practicality. Users end up having to recreate their itinerary during their trip.

02/


Trip planning as a group can be difficult, due to lack of collaboration, misaligned budget, and unmet expectations.

03/


Making an itinerary often requires using multiple apps, and tracking progress and information can be a big obstacle.

DEFINE

Persona

Problem Framing

User Journey

I identified 4 phases of travel planning, Define, Research, Commit, and Getting ready. By breaking down the tasks in each phase, I was able to identify and prioritise the opportunities that my design could address later.

User Journey

I identified 4 phases of travel planning, Define, Research, Commit, and Getting ready. By breaking down the tasks in each phase, I was able to identify and prioritise the opportunities that my design could address later.

Persona

I created the main persona based on user research, and used empathy map to understand their wants, behaviours, likes and dislikes.

Persona

I created the main persona based on user research, and used empathy map to understand their wants, behaviours, likes and dislikes.

Competitive Audit

Competitive Audit

Competitive Audit Summary:

Competitive Audit Summary:

There are many travel planner apps in the market, and users are having a hard time committing to one and make the most out of it, because none of the apps have a clear advantage.

There are many travel planner apps in the market, and users are having a hard time committing to one and make the most out of it, because none of the apps have a clear advantage, and users do not save time by using the same app(they have to repeat the process anyway).

  • The generated itineraries are in a lengthy text format which can deter users.

There are many travel planner apps in the market, and users are having a hard time committing to one and make the most out of it, because none of the apps have a clear advantage, and users do not save time by using the same app(they have to repeat the process anyway).

  • Users are using the AI-generated itineraries more as a reference than an actual itinerary.

There are many travel planner apps in the market, and users are having a hard time committing to one and make the most out of it, because none of the apps have a clear advantage, and users do not save time by using the same app(they have to repeat the process anyway).

  • Overly complicated use cases cause confusion and demotivate users.

There are many travel planner apps in the market, and users are having a hard time committing to one and make the most out of it, because none of the apps have a clear advantage, and users do not save time by using the same app(they have to repeat the process anyway).

Pivoted Problem Framing

Pivoted Problem Framing

Finding out the all-mighty google has a solution to people's travel planning isn't surprising, but even more surprisingly, google trips had been sunsetted in 2019.

There are many travel planner apps in the market, and users are having a hard time committing to one and make the most out of it, because none of the apps have a clear advantage, and users do not save time by using the same app(they have to repeat the process anyway).

Why did Google kill an app that is favoured by travellers? Google claims that they have moved most of the trips function to Google Map. However, based on my research, most google trips users are left clueless and disappointed by the replaced and missing functions.

Why did Google kill an app that is favoured by travellers? Google claims that they have moved most of the trips function to Google Map. However, based on my research, most google trips users are left clueless and disappointed by the replaced and missing functions.

Based on pain points I had discovered during research, I identified several unique opportunities that makes Google Trips(Or Google Travel) the potential one-stop platform that can help people make better itineraries with the help of AI, making travel planing easier than ever

Based on pain points I had discovered during research, I identified several unique opportunities that makes Google Trips(Or Google Travel) the potential one-stop platform that can help people make better itineraries with the help of AI, making travel planing easier than ever

How might we design a website interface redesign Google travel experience to increase confidence in AI-generated itineraries, while ensuring it is intuitive to use?


How might we design a website interface redesign Google travel experience to evoke confidence in AI-generated itineraries, ensuring they are intuitive and user-friendly?


How might we design a website interface redesign Google travel experience to evoke confidence in AI-generated itineraries, ensuring they are intuitive and user-friendly?

Design Principles


How might we design a website interface redesign Google travel experience to evoke confidence in AI-generated itineraries, ensuring they are intuitive and user-friendly?


How might we design a website interface redesign Google travel experience to evoke confidence in AI-generated itineraries, ensuring they are intuitive and user-friendly?

Finding a balance between AI and human input to optimise personalised experience.


How might we design a website interface redesign Google travel experience to evoke confidence in AI-generated itineraries, ensuring they are intuitive and user-friendly?


How might we design a website interface redesign Google travel experience to evoke confidence in AI-generated itineraries, ensuring they are intuitive and user-friendly?

Facilitate collaborative trip planning experience.


How might we design a website interface redesign Google travel experience to evoke confidence in AI-generated itineraries, ensuring they are intuitive and user-friendly?


How might we design a website interface redesign Google travel experience to evoke confidence in AI-generated itineraries, ensuring they are intuitive and user-friendly?

Break away from lengthy text, encourage users to interact with their itinerary.


How might we design a website interface redesign Google travel experience to evoke confidence in AI-generated itineraries, ensuring they are intuitive and user-friendly?


How might we design a website interface redesign Google travel experience to evoke confidence in AI-generated itineraries, ensuring they are intuitive and user-friendly?

Maintain consistency with Google's branding and user experience.


How might we design a website interface redesign Google travel experience to evoke confidence in AI-generated itineraries, ensuring they are intuitive and user-friendly?


How might we design a website interface redesign Google travel experience to evoke confidence in AI-generated itineraries, ensuring they are intuitive and user-friendly?

IDEATE

Initial Concept

Concept Evaluation

Potential Solutions

Through several brainstorming sessions, I came up with many ideas to address the problem statement, and chose some of them to prototype and test

Process

Through understanding user stories and insights from my research, I created 4 key user flows. I studied Google Travel's information architecture to better understand why people are missing the old Google Trip, and created an updated site map to incorporate the new user flows and to address the pain points.
Google has transitioned many of its trip functions to Google Maps. However, upon a heuristic evaluation, I identified that several vital itinerary creation features are dispersed across various Google applications, including Google Maps, Google Search, and Google Docs. To address this, I designed an updated site map that allows users to access these diverse features directly within the itinerary builder. This streamlined approach minimizes the need to juggle tasks across different platforms and offers users a clearer, more efficient experience.

Process

Through understanding user stories and insights from my research, I created 4 key user flows. I studied Google Travel's information architecture to better understand why people are missing the old Google Trip, and created an updated site map to incorporate the new user flows and to address the pain points.

Google has transitioned many of its trip functions to Google Maps. However, upon a heuristic evaluation, I identified that several vital itinerary creation features are dispersed across various Google applications, including Google Maps, Google Search, and Google Docs. To address this, I designed an updated site map that allows users to access these diverse features directly within the itinerary builder. This streamlined approach minimizes the need to juggle tasks across different platforms and offers users a clearer, more efficient experience.

  • User flow 01 - Generate itinerary outline

  • User flow 02 - Delegate a task to others

  • User flow 04 - Optimise Route

  • User flow 03 - Find and book a hotel

  • Site Map

PROTOTYPE & TESTS

Wireframe

Usability Test

High-fidelity Design

Wireframe

I decided to design the desktop MVP prototype first, in order to allow users to better access maps and booking information.

Wireframe

I decided to design the desktop MVP prototype first, in order to allow users to better access maps and booking information.

Usability Test

After I designed wireframes for the main user flows, I invited five participants for the usability tests using Quant-UX

Usability Test

After I designed wireframes for the main user flows, I invited five participants for the usability tests using Quant-UX

Each participants was asked to complete two tasks.

Task A: Generate a 4-day itinerary for Paris

Task B: Add hotel information to their itinerary

Findings

01/

Users tend to want to look at previous trips as a reference when creating new trips, thus making my trips another entry point for adding new trips

Users tend to want to look at previous trips as a reference when creating new trips, thus making my trips another entry point for adding new trips

Low fi screens

High fi screens

02/

Most participants were interested in editing their generated itinerary manually, and could not find the call to action which was previously located at the bottom of the itinerary.

Most participants were interested in editing their generated itinerary manually, and could not find the call to action which was previously located at the bottom of the itinerary.

Low fi screens

High fi screens

03/

Users want to be able to easily access itineraries by day, and adding places on the go.

Users want to be able to easily access itineraries by day, and adding places on the go.

Low fi screens

High fi screens

High-fidelity Prototype

High-fidelity Prototype

Collaborative itinerary made by friends and AI

Collaborative itinerary made by friends and AI

generated by AI chatbot, with flexibility and iteration built in mind.

Simple Integration

Simple Integration

powered by the Google ecosystem(Google Travel, Google Maps, google calendar, and Gmail). No more 100 opens tabs and dozens of email confirmations to keep track of.

Keep trip budget under control

Keep trip budget under control

Enabled by AI-predictive technology, get a reminder when it is the best time to book the cheapest fair.

Design System

Because this project is designed to be part of Google's ecosystem, I utilized Google's Material 3 Design library for style guide

Design System

Because this project is designed to be part of Google's ecosystem, I utilized Google's Material 3 Design library for style guide

Font

Font

Colour

Colour

Icon

Icon

Lesson Learned

Lesson Learned

While working with a problem, it is critical to approach the possible solution with an open mind. I started this case study with a biased solution in mind, but through research and iterations, I saw the flaws and biases within my own assumption, therefore pivoted to a solution that is true to research discoveries.

While working with a problem, it is critical to approach the possible solution with an open mind. I started this case study with a biased solution in mind, but through research and iterations, I saw the flaws and biases within my own assumption, therefore pivoted to a solution that is true to research discoveries.

Next Steps

Next Steps

While working with a problem, it is critical to approach the possible solution with an open mind. I started this case study with a biased solution in mind, but through research and iterations, I saw the flaws and biases within my own assumption, therefore pivoted to a solution that is true to research discoveries.

While working with a problem, it is critical to approach the possible solution with an open mind. I started this case study with a biased solution in mind, but through research and iterations, I saw the flaws and biases within my own assumption, therefore pivoted to a solution that is true to research discoveries.

Back to top

Hello@beijiagu.design
Always happy to chat about design and new opportunities

© 2023 by Beijia Gu. All rights reserved

Hello@beijiagu.design
Always happy to chat about design and new opportunities

© 2023 by Beijia Gu. All rights reserved

Hello@beijiagu.design
Always happy to chat about design and new opportunities:)

© 2023 by Beijia Gu. All rights reserved