Sentiment Analysis
Analyze and measure a speaker’s expressed feeling or enthusiasm about a specific topic using Sentiment API
Overview
High accuracy
The AI model behind the Sentiment API is designed to handle the complexities of human language and can accurately identify the sentiment of human conversations with high accuracy
Convenient use
Compared to other models, Symbl’s model limits the API calls you have to make if you require more than just sentiments increasing ease of use.
Aspect based sentiment analysis
Sentiment analysis performed in conjunction with topic modeling (through Topic API) so as to capture contextual relevance of the sentiment within topic as well as broader conversation
Real-time
Extract sentence level sentiment polarity to understand emotional context in mission critical conversations.
Multiple inputs
Sentiments are available for audio, video and text inputs.
Global polarity labels
Sentiment is provided as a polarity in the range -1 to +1 along with a suggested label from positive, neutral, and negative.
Use Cases
Customer Satisfaction
Understand speaker engagement, talk time, silence, and sentiment to analyze customer satisfaction and take action to improve outcomes of sales and customer care calls.
Agent Performance
Analyze the conversation dynamics of top performing customer care agents or sales reps and create a repeatable recipe for success.
Realtime Call Mitigation
Detect when calls are at-risk or customer satisfaction is falling and create agent-assist workflows to adapt the call-flow in real time.
Resources
Processing conversations
Start by processing real-time or pre-recorded conversations using the Streaming API and Async API.
Perform Topic level sentiment analysis
Identify members and topics in conversation and analyze messages and factors for each topic.
Set custom definitions
Fine-tune sentiment polarity labels and scores to cater to your unique business requirements.