AI Trust Score: A User-Centered Approach to Building, Designing, and Measuring the Success of Intelligent Workplace Features

This paper describes how to help product teams move increasingly towards responsible AI experiences by examining user trust within business contexts. We discuss the development, application, and validation of the AI Trust Score, which allows teams to measure the success of an AI feature from enterprise users’ perspectives. The Trust Score is a multi-dimensional metric which consists of several statements users respond to that inform their trust, including whether an AI feature helps with job efficiency and effectiveness and understanding how and when to use the feature for their job role. The metric, along with usability feedback, can be adapted to encourage UX-related conversations and actions within product or feature development processes.