I took the lead role in Concentration Coach by posting early in the project finder group. My current career in the AI realm of EdTech pushed me to lead a group in the project—a first for me, and something I look forward to developing further. Having an idea, brainstorming, and developing it is a key skill in my role as a product manager, and I wanted to take this opportunity to lead the team with an idea that aligns directly with my interests.
Challenges
Being in different time zones presented a challenge, as expected, and the group was only able to meet a handful of times, which wasn’t ideal. However, effective deadlines and visions were set at each meeting to ensure productivity. The time zone issue was made clear from the start, but next time, I would be more particular about my partners and conduct interviews to determine their skills, interests, and, most importantly, their availability.
Development
My role included student-side design. Work started as a GPT with a long prompt that was going to be tied to a Figma layout, but this wasn’t very user-friendly. Through my job, I discovered Claude’s ability, which led to feeding the prompt, along with detailed layout instructions, to achieve an MVP, which worked very well. Because Claude has a limited context window, I drew upon previous experience and tested out Lovable. This created the Concentration Coach seen today, allowing more functionality and development of other pages and modules.
The design was deliberate to make a simple UI that anyone could intuitively figure out, but I’m not sure much effort was put into testing it outside myself and Abdulehed, who made an account on Concentration Coach. Chris gave helpful suggestions about improving the layout, particularly the functionality of the timer. Nichole created an amazing design on Figma, and our vision came to life.
Tying to My Career
Currently, my company, TopSchool.ai, faces issues related to usability, particularly in the onboarding of teachers. They are expected to create a “TA” (context layer) that customizes their platform for better LLM outputs, but the results are quite poor. A large text box is a blank canvas to write with no guidance or “configuring the user” (Woolgar, 1990), and this leads to single sentences such as “I teach math” or “2 ADHD students, 1 ESL.”
This course has led me to rethink the entire process of how a TA is created, leading to a TA Builder Wizard with drop-downs, buttons, and checkboxes that focuses on the essentials of who the teacher is, rather than class-based information. Classes change too frequently for the old approach to effectively help a teacher over many years, so we are instead creating a teacher profile dedicated to stable (though editable) aspects of a teacher’s core, and this profile, combined with student context from another source, will create their TA. We have yet to implement this new approach, but I view it as “configuring the user” to create a highly effective TA the first time they log in to our platform.
References
Woolgar, S. (1990). Configuring the user: The case of usability trials. The Sociological Review, 38, 57–99.