Design for Chat and Voice
Digital Chatbot Development
Task
The Digital Marketing Director at Milwaukee School of Engineering (MSOE) identified a need for a more streamlined method of delivering event information to students on campus. Our team, comprised of myself as project lead and two of my peers, began a multi-week project lasting the entire semester.
Our task was to create a conversation flow for a potential MSOE event chatbot that would contain every possible interaction necessary to achieve the user’s end goal of finding a campus event. To achieve this goal, our team mapped out multiple conversation scenarios where the chatbot would have a positive, negative, or mistaken interaction to perfect these scenarios and ensure our user finds an event.
Takeaways
The client was pleased with our design, highlighting our chatbot’s creative name and method of achieving our use case. The client also noted the chatbot’s ability to reduce a conversation to minimal clicks rather than typing, drastically reducing the user's time-to-goal.
While I cannot provide a link to the functioning prototype, I can provide a link to our Figma Conversation Flow. Feel free to explore it!
Methods
To design the conversation flow as accurately and detailed as possible, we completed the following:
Conduct Market Research: To begin the project flow, we began with researching schools that currently have implemented chatbots for various use cases. We researched schools such as Purdue, University of California, Harvard, and more.
Design and Test Conversation Flow: After conducting research based on other schools and their chatbots, we analyzed their conversation flows to determine the best and worst practices. Based on these findings, we were able to create a customized conversation flow for our use case of finding events for MSOE students. Our team slowly tested it among our fellow peers to find inconsistencies and pain points which were promptly remedied.
Utilize Generative AI to Assist Designers: Under the guidance and encouragement of our professor, our team leveraged our knowledge and experience prompting generative AI like ChatGPT 3.5 to supplement our UX skills and generate content for our conversation flow.
Conduct User Interviews: Much like when our team tested conversation flows, we field tested our prototype chatbot (now affectionately named Remy Raider) with students around campus. The results were positive, and we moved forward with our design in our chosen chatbot builder, Tidio.
Remy Raider: Your MSOE Friend to help find events on campus! (Our Chatbot Icon)
Final Tidio Workflow
Iterations of Conversation Flow
Final Conversation Flow