Communication Customization via Sentiment Profile

A method, comprising:

• creating, by a system comprising at least one processor, a sentiment profile corresponding to a user account based on interactions between the user account and an entity that is associated with the system, wherein the interactions comprise:
• first data indicating types of queries and responses by the user account contained in the interactions that satisfy a specified interactions criterion,
• second data indicating patterns in behavior associated with the user account contained in the interactions, or
• third data indicating respective contexts of the interactions;
• after creating the sentiment profile, conducting, by the system, a first support communication session with the user account via a first artificial intelligence virtual agent, wherein the first artificial intelligence virtual agent determines what to communicate with the user account based on the sentiment profile for the user account;
• and
• conducting, by the system, a second support communication session with the user account using a second artificial intelligence virtual agent, wherein the second artificial intelligence virtual agent determines what to communicate with the user account based on the sentiment profile for the user account, and wherein the first artificial intelligence virtual agent differs from the second artificial intelligence virtual agent.

A non-transitory computer-readable medium comprising instructions that, in response to execution, cause a system comprising at least one processor to perform operations, comprising:

• creating a sentiment model for a user account based on interaction data representing interactions based on user input received during the interactions via the user account and entity input received during the interactions via an entity that is associated with the system, wherein the interactions comprise:
• first data indicating types of at least one of queries or responses represented in the user input that satisfy a defined interactions criterion,
• second data indicating patterns in behavior associated with the user input, or third data indicating contexts of the interactions;
• after creating the sentiment model and based on the sentiment model, initiating a first communication session with the user account via a first virtual agent, wherein the first virtual agent is to conduct the first communication session based on the sentiment profile;
• and
• initiating a second communication session with the user account via a second virtual agent, wherein the second virtual agent is to conduct the second communication session based on the sentiment profile.