Frustrated or angry customers can be detrimental to your business. It’s critical to be able to intervene in these conversations or analyze them later to understand what went wrong and to optimize your bot’s performance.
Sentiment analysis enables your Comm100 chatbot to identify the tone and emotional state of every message a visitor sends in real-time. This functionality helps you identify opportunities of improvement in a bot conversation and determine when to involve human agents in chats that need attention. Sentiment analysis allows you to mitigate risks and delight customers by identifying and interacting with those who need your most attention.
Every message that a visitor sends to a chatbot can be classified on a scale of Very Negative to Very Positive. Thresholds for each sentiment score can be adjusted on a sliding scale to suit your unique sensitivity requirements.
You can set rules and conditions based on the frequency of a certain sentiment type, thus optimizing the way your bot functions and the experience it delivers. For instance, a rule could be built that escalates a bot conversation to an agent every time a visitor sends 2 messages with a very negative sentiment.
Track the number of messages with Negative or Very Negative sentiment types for each conversation in the live chat history section. Improve your bot over time by investigating and addressing any problems or issues that occurred in those interactions.