AI sentiment analysis
Sentiment analysis is a Viva Engage premium feature that aggregates data across Viva Engage conversations to surface trends. As part of the Audience Analytics, sentiment analysis helps leaders and their teams understand the types of conversations happening in near real-time, as well as where, when, and how leaders might best connect with their audiences.
I worked with a team of 2 product managers and a team of data scientists, AI experts and engineers to come up with the experience with a version powered by ACS (live now) and a more advanced version powered by LLM (coming soon).
How sentiment analysis works
Sentiment analysis is powered by Azure Cognitive Services (ACS) that aggregates posts and comments from storylines and public posts to show audience overall sentiment and themes on the Audience Analytics page.
Overall sentiment
ACS analyzes posts written by members of the leader’s audience and measure themes across all public posts and comments on storylines and in communities and returns an overall sentiment trend, plus detailed breakdowns of different sentiment compositions.
Sentiment themes
ACS opinion mining surfaces frequently discussed subjects within the leader’s audience and determine the total sentiment expressed toward that subject (positive, neutral, negative, or mixed).
Threads that mention these subjects and have the highest number of comments and reactions are also surfaced. Leaders and their teams can use themes to gain context and participate in conversations more easily.
