Bank of America Erica AI Assistant

Platform: iOS, Android

Erica is an artificial intelligence application that interacts with Bank of America mobile app users via voice and consists of machine learning capabilities that lives within the mobile app system.

Team:

1 UX Designer (me)

1 Visual Designer

2 Engineers

1 Machine Learning Developer

1 User Researcher

1 Systems Analyst

1 Product Manager

the challenge

Developing an AI application with machine learning capabilities that would be incorporated into the existing Bank of America's consumer mobile app.  This was the first phase of Erica.

research

A formal usability lab testing was conducted on the Erica prototype.  The purpose was to gather user feedback on their impressions of Erica and its functionality.  Users were asked to complete moderator guided tasks and answer usage questions.  There were 2 prototypes that we compared; Open design vs. chat design.  This was followed by a desirability exercise.

Usability Study: Erica Agile Study Phase One

Open Design

Desirability Exercise

Open Design

Chat Design

Chat Design

​In the desirability exercise, we asked participants to come up with terms that best exemplify the two designs.

wireframes

Input field states and flow

Handoff Flow: FICO

final designs

For Phase 1, the usability study informed the final designs.