UW Comm Lead Chatbot
The only chatbot that had been implemented in a program at the University of Washington
At-A-Glance
Decreasing manual emails by 50%
The implementation of the chatbot on the UW Comm Lead website has proven to be a huge success. Since its launch, it has led to a remarkable reduction in the need for manual email responses, resulting in significant time savings for Heather, the academic director, amounting to hundreds of hours.
This project won the Design Award at the Comm Lead Screen Summit 2022.
Stakeholder
Director,
Academic Services
Target Users
Prospective Students
Project Type
Launched Project
Timeline
03/2022-06/2022
Tools
Tidio
Voiceflow
Figma
Team
1 PM, 1 Content Designer, 2 UX Researchers, 2 UX Designers
My Role
As the ux designer on this project, I collaborated with my team to carry out a competitive analysis, ensuring the selection of the most effective tool within our budget constraints. Together, we designed and tested the conversation flow using Voiceflow and developed a style guide in Figma.
Problem Statement
How might we design a chatbot that efficiently guides prospective students to the information required for applying to the Comm Lead program?
The director of academic services, Heather, wishes to streamline her communication with program applicants and students by avoiding repetitive email responses to general questions. She would love to have a chatbot implemented on the website which navigate perspective students to useful information about the program.
Stakeholder
Academic Director, Heather
"I wish to spend more time assisting students with their unique issues and helping them succeed in Comm Lead."
Reduce email reply work load
Would be able to help student solve common questions
Foster interest in our program among prospective students
Aligned with the UW branding
Understand Users
Users had trouble finding information they were looking for on the webiste
To provide an effective solution, we decided to engage in targeted research with our prospective students. This research aims to identify the specific information they require and the challenges they face in accessing it, ensuring we tailor conversations to meet their needs precisely.
Research Methods
Google Analytics
FAQs Collection
Surveys
Competitive Analysis
Research Goals
Get demographic information to identify our user groups
The bounce rate in the past
Collected by the director and it helps us to understand
Understand the needs of users
Learn from their their experience with other college chatbots
Identify the most effective chatbot tool within budget constraints
China
India
From Other Countries
User
Demographics
68% users are from the United States, 32% are international users
3 out of 5 survey participants had trouble finding information they were looking for
Google Analytics data shows that approximately 32% of our website traffic comes from outside the U.S., highlighting the substantial international student representation in the CommLead community. Additionally, user surveys revealed differing concerns between international and domestic students, prompting us to develop two unique personas based on our findings.
Domestic perspective students usually concern about tuition & financial aid and program duration
Since many of the domestic students were already working professionals, they often needed to afford the tuition by themselves. They wished to figure out the possibility of being a part-time student while working to afford the tuition and living expenses.
Domestic Student
James
Concerns
How to afford this degree
Application process
Questions
"How can I apply financial aid?”
“Can I attend part-time? How long does it take to complete the degree?”
International perspective students usually concern about visa and employment opportunities in the United States
International students were usually struggle with their visa situations. 90% of the Comm Lead perspective students wanted to find a full time job at U.S. after graduation.
International Student
Lisa
Concerns
Longer time allowance to find a job after graduation
The total cost for tuition and living in the U.S.
Questions
"Do Comm Lead degrees all STEM degrees?”
“Do I need to take any English proficiency tests?”
“Can I work in the U.S. while I am studying in Comm Lead?”
From Idea to Execution
Aim to build a chatbot that offers an engaging and user-friendly experience, delivering valuable information through human-like conversations
Building conversation flows
We divided all contents into 5 categories and made decision trees, and worked along with content designers to craft the conversations.
The basic structure of chatbot (Old Version)
We initially divided the conversation flows for domestic and international students, aiming to cater to their unique needs. (However, we later discovered that this separation was unnecessary.) During the conversation building process, we faced frequent challenges in categorizing information and struggled with the redundancy of details across various categories.
For instance, we were uncertain if separating the content into 'requirements' and 'steps' would more effectively assist students in navigating the information.
‘Admission’ flow (Old Version)
Unfortunately, We realized that we should conduct a card sorting to better understand our user expectations until we finished our first prototype.
So for the next step, we want to test out our prototype and understand how users feel about the existing flows and make adjustments based on the testing results.
Align this chatbot with both UW and Comm Lead branding and set up the tone for conversations
The chatbot should strike a balance between professionalism and approachability, offering friendly support without becoming too informal and empowering users as it assists them."
Testing for Better Outcomes
Examined 3 core usability issues and discovered new opportunities
After developing the first prototype, we decided to collaborate with researchers to conduct user testing. This aimed to resolve the categorization issues we encountered and to identify usability problems. Our goal was to enhance the chatbot experience, making it smoother and more interactive.
“At the end, I select 'I can't find what I want‘ but I don't have any menu showed up, I guess that's the end of the conversation. Wish there's option to go back to main menu.”
-One Participant who wished to have an option to go back to the main menu
“I’d like to join some Comm Lead events to know more about the program and the community, but I don’t know where to find it on the Comm Lead website.”
-One Participant who wished to have an option to go back to the main menu
Design Solution
Redesigned the new IA and conversation flows for a better and smoother user experience with a more interactive Comm Lead chatbot
We redesigned the information architecture and eliminated the flow separation between international and domestic students, as suggested by the testing results. Additionally, we renamed the categories to make them more user-friendly and added an event calendar to keep students informed about upcoming events.
A new conversation flow structure
Now featuring an engaging and user-friendly chatbot, designed to help thousands of prospective students easily navigate through the information on the Comm Lead website
Easy Navigation
1.Change the categorizations
2.Multiple “back” routes
Guide to Useful Resources and Events
1.For questions that could not be simply answered, students will be guided to different resources on UW website
2.An event calendar to help student explore all kinds of activities CommLead have
More interactive experience
1.An engaging test to help perspective students to understand what do they need for applications
2.Slower replying speed to mimic a human-like conversation
Contact Information for real person and community experience
Perceptive students would be able to talk to the director directly and join Comm Lead online community
Future Step
Fix the bugs we discovered after launch
We need to launch this chatbot asap to support students during the apply season. So after the launch, I reflected on this project and I discover this two issues that I want to delve into more.
Currently, the chatbot supports a text function that allows users to input specific keywords to initiate actions or conversations. However, this feature's current sophistication level is limited in handling complex textual interactions. Moving forward, my aim is to test whether we should disable the text input option or train a more sophisticated one.
Developing a more persuasive chatbot that encourages prospective students to sign up for Comm Lead events and engage with the community, thereby enhancing the likelihood of their interest in attending the program