Challenge: As our users are getting started with our product, marking recommendations as Useful and Not Useful trains the AI of that users unique preferences over time. However, many users complained that this process took about a month to complete.

 

User Research: Our Co-Founder conducted research in AI modeling to learn more about the effects that transparency can have on a users engagement in AI and ML products. Ultimately, she found that the more opportunities we can provide to demonstrate the AI that is working behind the scenes and invite our users into that, the better.

​

Prototype: Our Founder and Co-Founder approached me to think about a potential design for this solution. I provided them with a prototype displaying this feedback mechanism that would identify topics tagged within articles that they found Useful, and requested user confirmation in order to add those topics to their profile as a priority in order to receive more recommendations like that one.