AI recommendation Cards + Dashboard

 
 
Screen Shot 2021-05-23 at 6.56.54 PM.png
 

Objective

A product team asked me to design a card UI to use in a pilot program for their new machine-learning algorithm. The purpose of UI was to show purposeful calls-to-action based on various data sources, such as users' insurance information. For example, a user could be provided with a recommendation to get their flu shot or join a smoking cessation program if their data indicated that they would be beneficial programs for their health.

I was also required to build a "card hub" where users could access their current, completed, and discarded recommendations.

This project was designed in Sketch and hosted on Invision.

Design Constraints

  • Net-new product.

  • The product team asked me to create an "MVP" product with room for future enhancements.

  • Limited data were available to reference or to use in design decisions.

  • The card designs had to be "fluid" and work in different areas of a portal.

  • The card designs had to work across two differently branded platforms.

  • Limit on headline and body description text.

  • Upon product launch, users would only see 2-3 recommendations until more programs were available several months later.

Process

  • I created an audit of card designs from other internal projects and websites to inform my design decisions.

  • I worked with the UX research team to test my designs with participants and incorporate their feedback. For example, users preferred cards with a few lines of descriptions with a headline versus a standalone headline.

  • I created an illustration standard, so the cards had a "recognizable" look to them.

  • I worked closely with the development team to ensure my design could scale properly at different responsive breakpoints.

Screen Shot 2021-05-23 at 9.50.00 PM.png

Outcome

  • There is noticeable engagement with card UI and program conversion.

  • Other teams have adopted recommendation cards to use within their product spaces.

  • "MVP" dashboard design has outgrown its scope. It needs to be updated to include card categories, filtering options, and information on why users are recommended specific programs, such as weight loss.