Edtech software with user research
Challenge
Creating the shortest path to connect Rice University with students How do you grab the attention of an 18-year-old? That’s the question Rice University came to Table XI to answer. The school had recently launched its Doerr Institute for New Leaders, and needed to show students all it could do.
Solution
As we dug into user research, we quickly realized we weren’t going to have any luck getting students to look up from their phones. So we took advantage of the problem. Together, we developed a chatbot that would put the opportunities right in students’ hands. User testing helped us shape the chatbot experience until it felt like second nature, not a chore. As students texted in details on their academics and interests, the chatbot fed the data into a backend program, which paired them up with personalized leadership opportunities. All of this was done with Rice’s IT team right by our side, so we knew for sure that the University would be able to continue using and updating the chatbot long after Table XI wrapped. Applying — and extending — an idea for matching users to good deeds. It was our work on a similar project that brought The Doerr Institute to Table XI. A couple of years ago, we built a platform for the nonprofit WomenOnCall to connect professional women to volunteer opportunities. The Doerr Institute figured we could apply the same thinking to help them engage students — but what they needed went a step further. Where WomenOnCall’s users were seeking volunteer opportunities out, the Doerr team had to find a way to connect with students who might not even know there were leadership opportunities available. To figure out how to do that, we started with research. The Doerr Institute staff knew their audience — they work on the same campus as the target users after all, and they had copious research on-hand. What they hadn’t done was apply the systematic, UX research tactics Table XI builds projects around. We couldn’t start building for users until we knew what they wanted. So we worked with the Doerr team to find the real user needs the fastest way we know how — with a Google Design Sprint. Using a Google Design Sprint to go from need to chatbot A Google Design Sprint lets us brainstorm, prototype and test an idea in five days. With it, we were able to work with Rice to develop an idea that met their users’ needs, using technology the team hadn’t yet thought of. We had two goals going in: Make the product dynamic enough to capture a 19-year-old’s interest, and build something Rice’s technical team could control and maintain at the end of the project. We spent the first two days of the design sprint learning the landscape and developing a solution. Rice had a mass of student data, so we had plenty of data points to move us quickly into the brainstorming stage. We knew right away that success depended on meeting students where they spend all their time — their phones. Because it’s already a natural environment for them, a mobile solution was an opportunity to catch them where they’d be more comfortable and forthcoming. An app didn’t feel right — it’d need to be built and maintained, and what would prompt a student to download it? But a chatbot would allow us to engage with students on their own terms (and on their unlimited SMS plans!) On the third day of the sprint, we started to visualize how it would work: After seeing a flyer on campus, or talking to a peer during Orientation week, students would text a code to a specific number, then begin interacting with the chatbot AI. The interactive chatbot would ask questions about the person’s major, their extracurriculars, and what they felt passionate about in their daily lives. Then it would use this information to put together a custom program of leadership opportunities based on the students’ interests. To make it simple for them to participate, the chatbot would send a link to a responsive webpage with all the program information, so when students felt ready to take action, they could do it right from their phone. We built a working prototype of this idea on the fourth day of the design sprint, and on the fifth, we tested it on representative users. The student feedback validated the idea, and started to give us a sense of what response speed would feel natural and welcoming. Then we had to tackle the hardest part, making the chatbot talk like a normal person. “One of Table XI’s biggest contributions was figuring out that this needed to be an SMS-driven process. Students don’t like to go to websites because it’s cumbersome and requires a laptop. They don’t like apps because it’s too much of a commitment to download one.” Tom Kolditz, founding director of the Doerr Institute for New Leaders at Rice University Building a chatbot that sounds like a person — but not like your mom We have never felt older than we did when we tried to text like a teenager. While our technology underpinned the whole project, all users experienced was a series of texts. The chatbot’s success hinged on something intangible: Did it feel like a robot, or did it feel like a human? Rice University students needed to want to engage with the chatbot for the tool to be effective. We had to make an intelligent chatbot, and we had to make sure it didn’t feel like one. Our first solution was asking students’ names right after they texted the code. That way, later on, the chatbot chat would include the person’s name in certain responses, the same way a friend’s text would. Not all of our human touches worked. Getting the slang right was painful, and as it turns out, unwelcome. We thought students wanted an experience that mimicked talking to a person, but they told us they were perfectly comfortable talking to an AI. One of our first messages used the phrase “BRB.” The students flinched, telling us it was like communicating with a mom trying a little too hard to be cool. They responded well to emojis but didn’t want us to overdo it trying to sound personable. With the backend we build, Rice is able to take this feedback and respond on the fly, honing in on a tone that’s balanced enough to provide people with trustworthy information while still feeling like a conversation. Working closely with a client’s developers to create chatbot tools that outlast Table XI One of our two goals from the start was to build a product Rice could takeover. The Doerr Institute hired us because they needed help solidifying an idea and putting it into practice, quickly. But Rice University already has its own systems and its own team of developers. We had the chatbot itself — a custom Rails application built around communications platform Twilio — plus a website for displaying those personalized leadership recommendations and a backend where Doerr staff could update the chatbot’s logic and add recommended programs. We worked with the Rice IT team to develop the code and QA standards we employ in other projects, so they could persist that level of quality after the project ended. That allowed Rice to own the product, so the chatbot could evolve without Table XI’s involvement. Giving the Rice team keys to everything we were building gave them the same sense of ownership Table XI had. They were involved in every part of the project, from ideation to execution, so they were fully invested in the results. Raising Doerr’s visibility across campus with a UX strategy that meets students where they are Getting all this student engagement just right was necessary, because the Doerr Institute was brand new. We couldn’t count on students to know what it was, or what it could do for them. When Rice first reached out, they made it clear their purpose was to build a useful tool and market the Institute’s new student-focused resources. That’s why we worked so hard to make the chatbot incredibly easy to use — in a way, it was an ambassador for the entire Doerr Institute. The chatbot let us reach students where they live. The Doerr Institute was able to communicate its value and essentially start conversations with students all over campus. Now they can text with students to learn what they care about — and adapt their programming to suit — without the experience ever feeling forced.
Want fresh marketing stories in your inbox?
Get the weekly email that makes you a marketing genius.
Stay informed and inspired, for free.