Answer time prediction in a contact center service of a provider network

A computer-implemented method for answer time prediction in a provider network, the method comprising:

• receiving, at the provider network, a set of contact queuing context-actual answer time data for a set of contact inquiries serviced as a first set of contact queuing context-actual answer time data, wherein a set of answer times were predicted for the set of contact inquiries based on a previous version of a queuing model, and wherein the previous version of the queuing model was learned based on a previous set of contact queuing context-actual answer time data for a previous set of contact inquires serviced, wherein the provider network is implemented by one or more electronic devices;
• learning, at the provider network, a new queuing model based on the first set of contact queuing context-actual answer time data and the previous set of contact queuing context-actual answer time data for the previous set of contact inquiries serviced or the previous version of the queuing model;
• receiving, at the provider network, a request to predict an answer time for a target contact queuing context;
• predicting, at the provider network, an answer time for the target contact queuing context based on the new queuing model;
• and
• providing, by the provider network, the answer time prediction to an electronic device that sent the request to predict an answer time for the target contact queuing context.

A system comprising:

• a first one or more electronic devices to implement a contact center service in a provider network, the contact center service including instructions that upon execution cause the contact center service to:
• receive, while a delayed processing timing has not been met, a set of contact queuing context-actual answer time data for a set of contact inquiries serviced as a first set of contact queuing context-actual answer time data, wherein a set of answer times were predicted for the set of contact inquiries based on a previous version of a queuing model, and wherein the previous version of the queuing model was learned based on a previous set of contact queuing context-actual answer time data for a previous set of contact inquires serviced;
• and
• provide, via a contact communications channel, an answer time prediction for a target contact queuing context based on a new queuing model;
• a second one or more electronic devices to implement a machine learning training service in the provider network, the machine learning training service including instructions that upon execution cause the machine learning training service to:
• learn, when the delayed processing timing has been met, the new queuing model based on the first set of contact queuing context-actual answer time data and the previous set of contact queuing context-actual answer time data for the previous set of contact inquiries serviced or the previous version of the queuing model;
• a third one or more electronic devices to implement a machine learning inference service in the provider network, the machine learning inference service including instructions that upon execution cause the machine learning inference service to:
• receive a request to predict the answer time for the target contact queuing context;
• and
• predict the answer time for the target contact queuing context based on the new queuing model.