Our project started with a brief from DuoLingo which asked our team to create a new way for their users to connect. We were also made aware of the app’s existing limitations with voice-recognition and AI chat-bots, but were encouraged to openly consider designing with advanced technologies in mind.
Our team developed a research plan and we used a range of methods, including:
competitive analysis, user interviews, affinity mapping,
persona creation, user flows, wireframing, prototyping,
usability testing, and usability test surveys.
Our research identified that users felt nervous about practicing their language skills in real life. They wanted a way to practice conversational skills with other users in a safe, low-stakes environment. We also found that users enjoyed a variety of resources and games in DuoLingo, and that they wanted their language learning to translate to real-life scenarios.
After compiling our research from interviews and the affinity map we were able to hone in on our users frustrations and Goals. This helped us create our ideal user named Ella Watts
I sketched some ideas for a user flow that might illustrate Ella’s path through one of these possible solutions, and shared and discussed our lo-fi wireframes as a group. After voting on what elements we liked most from each person’s sketches, then I put together a mid-fi prototype.
I developed an interface for DuoScenarios, a conversational immersion game for DuoLingo, to help users like Ella gain confidence in their conversational skills through game-based learning.
When DuoScenarios is opened, users are prompted with the first scenario and matched with a partner from their unit. Duo will guide the conversation to ensure it stays within the constraints of the unit, and partners take turns recording sentences and responding to each other. This interface provides a foundation for testing, and more details about the design are included below.
Even though our mid fidelity prototype didn't do so well, we found some easy-to-fix problems. So, we got to work and made the necessary changes pretty quickly. After that, we tested it and got some promising results. You can check out the improved prototype here to see how we did it.
View Prototype