Abstract: Techniques for integrating third-party analytics with virtual-assistant enabled applications are disclosed. A third-party analytics service trains a machine learning model, using labeled training data including (a) phrases corresponding to sales offers made to consumers and (b) sales conversion outcomes associated with the phrases. The service receives, from a consumer-facing application, a user query submitted via a virtual assistant interface. The service applies the user query to the machine learning model, to obtain a recommended phrase for the consumer-facing application to use in response to the user query. The recommended phrase is: based on one or more of the phrases used to train the machine learning model; responsive to the user query; and based on a likelihood of achieving a sales objective associated with the consumer-facing application.

Abstract: Techniques for providing third-party analytics via a virtual assistant interface are disclosed. A third-party analytics service trains a machine learning model, based at least on interaction histories of users of a consumer-facing application. The interaction histories include sales data associated with the users. The third-party analytics service receives, via a virtual assistant interface, a request for a recommended marketing strategy to be targeted at one or more users of the consumer-facing application. The third-party analytics service applies the request to the machine learning model, to obtain the recommended marketing strategy responsive to the request. The recommended marketing strategy is based at least on a predicted effectiveness of the recommended marketing strategy targeted at the one or more users of the consumer-facing application. The third-party analytics service presents, via the virtual assistant interface, the recommended marketing strategy responsive to the request.

Abstract: Techniques for integrating third-party analytics with virtual-assistant enabled applications are disclosed. A third-party analytics service trains a machine learning model, using labeled training data including (a) phrases corresponding to sales offers made to consumers and (b) sales conversion outcomes associated with the phrases. The service receives, from a consumer-facing application, a user query submitted via a virtual assistant interface. The service applies the user query to the machine learning model, to obtain a recommended phrase for the consumer-facing application to use in response to the user query. The recommended phrase is: based on one or more of the phrases used to train the machine learning model; responsive to the user query; and based on a likelihood of achieving a sales objective associated with the consumer-facing application.

Abstract: Techniques for providing third-party analytics via a virtual assistant interface are disclosed. A third-party analytics service trains a machine learning model, based at least on interaction histories of users of a consumer-facing application. The interaction histories include sales data associated with the users. The third-party analytics service receives, via a virtual assistant interface, a request for a recommended marketing strategy to be targeted at one or more users of the consumer-facing application. The third-party analytics service applies the request to the machine learning model, to obtain the recommended marketing strategy responsive to the request. The recommended marketing strategy is based at least on a predicted effectiveness of the recommended marketing strategy targeted at the one or more users of the consumer-facing application. The third-party analytics service presents, via the virtual assistant interface, the recommended marketing strategy responsive to the request.