Sentiment training


WhosOn comes complete with a sentiment analysis algorithm. This is used for two things: to indicate chat mood in real-time, and to produce an overall sentiment score after the chat has finished.

To make this feature as powerful as possible, WhosOn uses ongoing sentiment training. The more data it has, the smarter its analysis. Here’s what you need to know about sentiment training:

• WhosOn recognises sentiment within chats, and displays a visual mood indicator

• This real-time mood reflection will change over the course of the conversation

• At the end of the chat, a definitive sentiment score is calculated

• This final scoring trains the algorithm by cross-referencing the customer’s chat rating

• As more chats and ratings are completed, the scoring becomes more accurate over time