System, method, and computer program for preempting bill related workload in a call-center

A non-transitory computer readable medium storing computer code executable by a processor to perform a method comprising:

• collecting a first dataset including a plurality of records, wherein the first dataset includes:
• a plurality of customer records,
• a plurality of billing records for each of the customers, wherein each billing record includes at least one billing item, and
• a plurality of call incident records, wherein each call incident record includes:
• an identification of the calling customer,
• an identification of a billing record, and
• an identification of a billing item discussed with the calling customer during a past call with a call-center;
• training a first artificial intelligence model (AI-model), using the first dataset, to recognize at least one pair of a first customer type and a first billing item type, and an associated first probability that such pair results in a call to the call-center;
• collecting a second dataset comprising new billing records;
• using the first AI-model to detect at least one billing record in the second dataset having a probability higher than a predefined threshold probability that the customer associated with the billing record will call the call-center;
• collecting a third dataset including:
• customer identification,
• a billing item discussed with a customer, and
• a resolution accepted by the customer to resolve a problem associated with
• the billing item;
• training a second AI-model, using the third dataset, to recognize at least one triplet of a second customer type, a second billing item type, and a resolution type, and an associated second probability that a resolution of the resolution type is accepted by the customer;
• for each billing record of the at least one billing record in the second dataset having a probability higher than a predefined threshold probability that the customer associated with the billing record will call the call-center:
• using the second AI model to determine, for a billing item in the billing record, at least one resolution for billing item, and
• selecting a resolution from the at least one resolution determined for the billing item, and
• performing the selected resolution to preempt the call to the call-center,
• wherein the selected resolution is performed before a bill including the billing item is issued to the customer.

A system, comprising:

• a non-transitory memory storing instructions;
• and
• one or more processors in communication with the non-transitory memory that execute the instructions to perform a method comprising:
• collecting a first dataset including a plurality of records, wherein the first dataset includes:
• a plurality of customer records,
• a plurality of billing records for each of the customers, wherein each billing record includes at least one billing item, and
• a plurality of call incident records, wherein each call incident record includes:
• an identification of the calling customer,
• an identification of a billing record, and
• an identification of a billing item discussed with the calling customer during a past call with a call-center;
• training a first artificial intelligence model (AI-model), using the first dataset, to recognize at least one pair of a first customer type and a first billing item type, and an associated first probability that such pair results in a call to the call-center;
• collecting a second dataset comprising new billing records;
• using the first AI-model to detect at least one billing record in the second dataset having a probability higher than a predefined threshold probability that the customer associated with the billing record will call the call-center;
• collecting a third dataset including:
• customer identification,
• a billing item discussed with a customer, and
• a resolution accepted by the customer to resolve a problem associated with
• the billing item;
• training a second AI-model, using the third dataset, to recognize at least one triplet of a second customer type, a second billing item type, and a resolution type, and an associated second probability that a resolution of the resolution type is accepted by the customer;
• for each billing record of the at least one billing record in the second dataset having a probability higher than a predefined threshold probability that the customer associated with the billing record will call the call-center:
• using the second AI model to determine, for a billing item in the billing record, at least one resolution for billing item, and
• selecting a resolution from the at least one resolution determined for the billing item, and
• performing the selected resolution to preempt the call to the call-center,
• wherein the selected resolution is performed before a bill including the billing item is issued to the customer.