A group of interdisciplinary teammates and I worked over the course of a month to create a conceptual product that utilizes AI in a meaningful way. Our team chose to create a platform for users seeking healthcare from a new provider. Current solutions demand much from the end user, only to provide them with a large amount of results that may not be relevant to their needs.
Often, a patient may experience a standard symptom such as pain or weakness and then pushed off by the doctor, something that women and minority communities happen all too well. By asking the patient to describe their symptoms and pairing what we already know about the patient, we can better target specific doctors that may be a better fit for our particular patient. Our tool would:
We can utilize information provided by the user's employer to cut out repetitive steps that leaves the user asking "shouldn't they already know this?"
Sorting by distance is great, but a distance of half a mile or 5 miles can mean the same thing to someone who relies on public transportation or is traveling from the suburbs into the city to see a specialist.
We made a decision to show the user the maximum number of results they could receive from the provider's network, but then to show the number going down as the results narrowed. This lets the user know the value of their insurance provider's network, while also showing that the platform is doing work in real-time to generate personalized results for the user.
Finally, the platform shows the user their results, highlighting the doctor that is most recommended and why. Of course, users will want to see what other options are out there, so we provide a set of top matches to compare.
There are ethical questions that we chose to not thread with this exploration. While the end product is an improvement as is, we did not flex the AI functionality as much as we could have, but doing so would require to ask ethical questions that were frankly, out of scope. Could we allow a user to give us a detail their symptoms and with machine learning, ask follow up questions that may either get to underlying issues or allow us to direct the user to a more specific practitioner that specializes in the problem we identified?
The concept has plenty of real world application. What if instead of instead of healthcare insurance providers, this was a product for Autozone that empowered users to diagnose and repair their own car issues?