IBM watson // 2017
A tool that provides students confidence and easy navigation during the pre-registration process
While enrolled in IBM Design Studio’s Radical Collaboration, I worked with a team of six to integrate Watson’s technology in higher education – specifically class registration at our shared university. With the mentorship of IBM’s senior designers, I was able to own the problem through user research, UX, and visual design. At the final presentation, my team provided a solution to alleviate stress during the pre-registration process, accompanied by a structured prototype for the reimagined experience.
Equipped with technology that’s easily integratable with any API, IBM Watson trains itself through machine learning algorithms and repeated use. The computer system proved its ability to process colloquial language when it starred on Jeopardy and took home the first place prize of $1 million.
IBM Watson
→ Can process natural language
→ Ability to aggregate data and intelligently retrieve it
→ Smarter functionality with increased use
CLASS REGISTRATION: A STRESSFUL EXPERIENCE
As current students at The University of Texas at Austin, my team and I had our fair share of painful registration experiences. The process depended on an online input system that left no room for error, and since required classes were seldom available, students spent hours preparing multiple backup plans. In addition, most students were dependent on meetings with advisors that lacked specification and therefore, lacked guidance. This made damage control an inevitable part of the student registration experience – something we wanted to help alleviate with Watson’s technology.
With a potential user pool of 50,000 students, I conducted six weeks of design research and iteration on our university campus. I wanted a better understanding of the problem and as-in scenario, so I specifically measured how students’ confidence shifted during the process and the impact of various registration tools.
FOCUSING ON PRE-REGISTRATION
I discovered that students’ confidence before registration had a high impact on their experience day-of. In several interviews, users expressed regret for not preparing enough, regardless of how much time they spent researching beforehand. Even seniors, who had years of practice, expressed frustrations about the lack of support they felt leading up to registration day.
I noticed that on average, students spent around 3 hours allocating data from various sources to create their schedules for the next semester. This didn’t include meeting with advisors for common problems like fixing registration restrictions, or bars. How can Watson help make this pre-registration time more efficient and individualized?
UNDERSTANDING DIFFERENT NEEDS
It was important for me to understand the general types of registration experiences by mapping out student personas with the team.
I noticed that on average, students spent around 3 hours allocating data from various sources to create their schedules for the next semester. This didn’t include meeting with advisors for common problems like fixing registration restrictions, or bars. How can Watson help make this pre-registration time more efficient and individualized?
The common thread between our user personas was their need for personalized assistance. These students spend hours going back and forth between different tools while preparing their schedules, which a) wastes time and b) overwhelms them with information.
DESIGNING TO REDUCE CLUTTER
The goal was to create a digital extension that would cut down time spent looking for general public information, such as degree requirements and bars. This way, students are able to focus on getting help on more specific matters, such as adding minors and finding courses that meet at the intersection of their multiple interests. This was done by integrating IBM Watson’s natural language and data compiling capabilities into a registration chatbot messenger. By providing students with one tool that addresses several of their needs, we hoped to provide students with confidence in their decisions and therefore reduce their stress levels.
One major feedback we received was the user’s desire to compile given information into categories. We sought inspiration from Pinterest’s familiar and easy-to-use folder making system and integrated a way for students to “favorite” and gather their desired classes in one place. This also gave way to a level of personalization that students enjoyed.
ITERATING WATSON ADVISOR
Through multiple user testing phases, I learned that our users:
→ Felt comfortable using natural language after being instructed to do so, but were often very explicit and tried to get a perfect answer out of one message
→ Had the urge to interact with anything displayed.
→ Wanted as much information as possible displayed from a single response.
→ Desired the ability to store and organize the information they were given
OUR SOLUTION
Our solution aimed to provide an aggregated data platform for students to seek answers for individualized questions about registration. We wanted to minimize the time students spent during the pre-registration stage and lessen the confusion that students felt when looking for information online. This resulted in a personalized chatbot that processes natural language, intelligently retrieves data from multiple online registration tools, and becomes smarter functionally with increased use. Due to Watson’s ability to train with any API, the product would be easily scalable for other higher education institutions.
Through increased use, this chatbot will better help students navigate their degree plan and class requirements before they head into registration.
REFLECTION
→ Design with Research When designing with advanced technology like IBM Watson, it was easy for me to focus on the user’s final interactions with the product. But as I progressed through the research phase, I realized how important it is to design for specific moments during the user’s journey. These moments provide valuable insight on users’ emotional journeys that in turn, help create an impactful product.
