JumpStartCSR
Aiding Athletic trainers with AI
Project Type
Capstone Project
Role
VR Developer & Designer
Timeline
10 Weeks
Team Size
4 Team Members
JumpStartCSR is a Seattle based company that uses their artificial intelligence, called “Holmz,” to make recommendations about a person’s health. It uses biometric data obtained from wearable technology such as smartwatches to guide these recommendations.
JumpStartCSR wanted us to speak to athletic training staff and see how exactly the Holmz AI could be integrated within their workflow. We found participants, interviewed them, recorded our findings, and moved into the design phase of the project. Because this was an exploratory part of the research process, we ended up cold-calling/emailing over 200 potential participants..
We chose these research questions in order to fulfill the requirements of our stakeholders, as well as satisfy the needs of the target audience for the product.
We then began searching for potential testing candidates according to the following criteria:
Once we decided on the inclusion/exclusion criteria, we cold-called/emailed over 150 potential participants, went in person to the athletic offices at UW, and asked for any connections that JumpStartCSR could provide us. Due to privacy concerns, JumpStartCSR only provided us with about 5 contacts, of which only a couple responded.
We ultimately landed five participants that fit our inclusion criteria, spanning specific job titles from Athletic Trainers to Athletic Directors.
Our interviews aimed to understand the current workflows, processes, challenges, and other opportunities for the Holmz AI to assist Athletic Trainers (ATs) in monitoring athlete health. To accomplish this, our group adopted a semi-structured interview approach, where two team members worked together to interview the participants. Interviewees were chosen according to the inclusion-exclusion criteria, and valuable insights were gained, specifically regarding the processes for evaluating athlete readiness to play.
Using the research questions written above, we developed a list of interview questions and moderator's guide, linked here.
We found multiple takeaways from our interviews, but the following three were the most impactful.
There are standardized mobility and screening tests that most ATs use to gauge the health of different musculoskeletal groups.
These tests are done to determine an athlete’s readiness to return to play, as well as assess the likelihood of an injury to specific parts of the body as well as identify pre-existing conditions or injuries that may impact an athlete’s health.
The participants were very interested in the sleep, fuel, and hydration habits of their athletes.
Each participant who mentioned it spoke to how important these three aspects of an athlete’s life are to their ability to perform when healthy, as well as their speed of recovery when injured. Further, one trainer we spoke to made it clear that if these aspects of health are ignored, other recovery methods, like massages, are significantly diminished in positive impact for recovery.
Multiple participants made sure to mention how unreliable self reported data can be, especially in the context of health assessment and adherence to prescribed plans.
Athletes have to self report certain measures, like pain. This is unreliable because measures such as pain are subjective; they change from person to person. An AT participant commented that athletes are also sometimes deceitful and don’t always tell the truth about their personal habits, or adherence to any prescribed plans. This creates a clear disconnect between the trainer and athlete and makes the ATs job harder because they have to work under false information.
After assessing our takeaways from the interview process, we decided to finalize a focus on Athletic Trainers, while aiming to design in a way that would still empower coaches and other sports professionals, as well as serve as a bridge for ATs to communicate any data they needed to them using hard and set metrics.
We made this decision due to the opportunity to use AI to help reduce ATs time crunch as well as to help improve the ease of access that they would have to quick and accessible metrics. The desire for information on adherence as well as sleep-fuel-hydration were core needs that we believed we could match, especially in a way that caters to the AT demographic’s needs.
At the start of the design process, our team held multiple rigorous whiteboard sessions to plan for the design of the User Interface. We wanted to make sure that we had a plan with strong foundations in our research to ensure we wouldn’t run into problems down the line.
We compiled our findings and began brainstorming ways we could cater to the needs of the potential users we interviewed:
At the start of the design process, our team held multiple rigorous whiteboard sessions to plan for the design of the User Interface. We wanted to make sure that we had a plan with strong foundations in our research to ensure we wouldn’t run into problems down the line.
We compiled our findings and began brainstorming ways we could cater to the needs of the potential users we interviewed:
At the start of the design process, our team held multiple rigorous whiteboard sessions to plan for the design of the User Interface. We wanted to make sure that we had a plan with strong foundations in our research to ensure we wouldn’t run into problems down the line. Presented are some sketched out frames for the edit plan feature of the application.
Now we were ready to start working in Figma. After tinkering more with the information architecture, we realized that some things needed to be added, most notably what was included in the hotbar at the bottom of the page to help with ease of navigation.
At this point, we were comfortable with where our project was, and we were ready to start making our final deliverable.
Throughout the development of the high-fidelity prototype, we had four user testing sessions. Of the user testing sessions we had, we only had one major suggestion, being that the dashboard include more aspects of the calendar. This served as a “quality of life” improvement from a housekeeping standpoint for ATs that would use this product.
Finding a way to help market a brand new product to people who don’t think they need one.
Design is problem solving –for both the company and the user. The company has a product that they want to sell to some demographic, but how do you sell that solution if the demographic doesn’t think they have that problem?
By conducting user research, we were able to uncover areas in the workflow of Athletic Trainers that could be improved by leveraging the HOLMZ AI.
Administrative and housekeeping tasks are a bigger deal than we thought.
A ubiquitous piece of feedback received during user testing was that there needed to be more “housekeeping” elements available on the dashboard. Specifically relating to the project, this meant that we needed to include critical information such as the daily agenda, appointment times, and relevant athlete profiles for that day directly on the dashboard.
AI has never, and likely never will be a replacement for human Athletic Trainers.
We were making an aid to help the athletic trainers, not an app that does their work for them. By making this service work more in conjunction with Athletic Trainers, as opposed to working "for" them, we were able to make a product that fit both the company and prospective customer's needs.