Yes or No Results

For classifications, also known as file-based events, VoiceBase will identify if a distinct event occurred and will return a positive or negative result. VoiceBase uses big data techniques to build a model and then predict the presence of any signal: Was a sale made? Is this customer a ‘hot lead’? Was the agent polite? These types of events are defined as call classifications.

Yes or No Results

For classifications, also known as file-based events, VoiceBase will identify if a distinct event occurred and will return a positive or negative result. VoiceBase uses big data techniques to build a model and then predict the presence of any signal: Was a sale made? Is this customer a ‘hot lead’? Was the agent polite? These types of events are defined as call classifications.


Customized vs. General

VoiceBase provides a suite of general classifications for many common use cases such as ‘Appointment Made,’ or ‘Upset Customer.’ Customers can provide pre-tagged calls as training data to create a custom classification or simply to improve accuracy for their specific use case.

Customized vs. General

VoiceBase provides a suite of general classifications for many common use cases such as ‘Appointment Made,’ or ‘Upset Customer.’ Customers can provide pre-tagged calls as training data to create a custom classification or simply to improve accuracy for their specific use case.


See Results

The JSON response for call classifications, or event-based predictions, includes the ModelID that was run to create the result, the type of predictor, and the actual result. The ModelID is a unique ID we give to you after building your prediction model that you are able to use to reference the model in API calls. The type will either be ‘binary’ or ‘multiclass’ for an event-based predictor. Binary indicates there are only 2 result possibilites, multiclass indicates there are many result possibilities. The actual result depends on how the model was built. For example, a binary predictor for hot lead detection might have the results “hot lead” and “non-lead”, a multiclass predictor for call categorization might have the results “customer”, “vendor”, “spam”, and “partner.”