System and method for managing routing of customer calls to agents
摘要
說明書
In an embodiment, customer demographic data also includes data using zip-level features of the system, which provides a coarser representation in building the predictive model. Such zip-level features employ variables that have resolution at the zip-level for each individual in the zip code. In an illustrative embodiment, zip-level data for individual income is associated with a median value of income for each individual in the zip code. Reasons for using zip-level data in predictive modeling include, for example, lack of a statistically significant difference in model performance as a function of any polymr match score threshold; simplicity of collecting only the name and zip code in the telegreeter process; and privacy considerations as to individual-level data.
In various embodiments embodiment, in predictive modeling of inbound callers, inbound queue management system 402 uses a fast-lookup tool (e.g., polymr) that analyzes customer identifiers of inbound callers in real time to retrieve customer data, such as customer demographic data, matched to the customer identifiers. In an embodiment, the polymr fast-lookup tool is a lightweight, extensible search engine, or API, implemented in the Python object-oriented programming language, https://www.python.org/. In various embodiments, the polymr tool performs real time matching of data in the customer demographic database 432 to a customer identifier for a given lead. In various embodiments, as a preliminary to using data in real-time predictive modeling of inbound callers, inbound queue management system 402 indexes the data by applying the search engine to customer identifiers in customer training data, and stores this index as an internal enterprise database 420.
