obtaining a message in a session between a user and an electronic conversational agent, the avatar being a visual representation of the electronic conversational agent;

performing sequential motion parsing to determine a facial animation based on at least one part of the message, wherein the sequential motion parsing determines a time period of the facial animation and the sequential motion parsing is based on at least one of: a keyword in the message, an emotion of the message, a sentence type of the message, a pronoun in the message, and a greeting in the message;

presenting the facial animation, the facial animation having a first facial expression associated with first content of the message and an instant variable second facial expression interspersed within the first facial expression during the presentation of the facial animation, the instant variable second facial expression being variable and different from the first facial expression based on second content of the message that is different from the first content, the instant variable second facial expression being created in real time and interspersed within the first facial expression in real time during the presentation of the facial animation, the message including a waiting period that allows for complete presentation of the instant variable second facial expression, wherein the first facial expression is presented after the complete presentation of the instant variable second facial expression; and

generating at least one motion of the avatar based at least on the facial animation.