US 12,271,930 B2
Method and apparatus for automated inter-account interactions
Angelo D'Alessandro, Studio City, CA (US)
Assigned to LIVEPERSON, INC., New York, NY (US)
Filed by LIVEPERSON, INC., New York, NY (US)
Filed on May 18, 2023, as Appl. No. 18/198,923.
Application 18/198,923 is a continuation of application No. 17/537,109, filed on Nov. 29, 2021, abandoned.
Claims priority of provisional application 63/119,264, filed on Nov. 30, 2020.
Prior Publication US 2023/0289850 A1, Sep. 14, 2023
Int. Cl. G06Q 30/0279 (2023.01); G06N 20/00 (2019.01); G06Q 10/1093 (2023.01); G06Q 20/02 (2012.01); G06Q 20/10 (2012.01); G06Q 20/12 (2012.01); G06Q 20/40 (2012.01); G06Q 30/01 (2023.01)
CPC G06Q 30/0279 (2013.01) [G06N 20/00 (2019.01); G06Q 10/1093 (2013.01); G06Q 20/02 (2013.01); G06Q 20/10 (2013.01); G06Q 20/12 (2013.01); G06Q 20/40 (2013.01)] 51 Claims
OG exemplary drawing
 
1. A method of automated interaction, the method comprising:
training a neural network using training data, wherein the training data is associated with historical user selections for historical interactions between pairs of users, wherein training includes setting one or more numeric weights of the neural network, and wherein the one or more numeric weights are associated with respective connections between respective pairs of nodes in the trained neural network;
receiving information about a plurality of accounts and corresponding calendars, wherein the plurality of accounts include a user account associated with a user calendar and an agent account associated with an agent calendar, wherein the information identifies associations between the plurality of accounts and a plurality of categories of goods;
receiving an indication requesting an interaction to involve a user associated with the user account;
dynamically analyzing the information using the trained neural network in real-time as the information continues to be received, wherein the trained neural network selects an agent associated with the agent account for the interaction based on the analysis indicating that the user calendar and the agent calendar have matching availability for a calendar event and that the user account and the agent account are both associated with a specific category of goods;
scheduling the interaction between the user and the agent for the calendar event;
initiating the interaction between the user and the agent as scheduled, wherein the interaction is associated with the specific category of goods;
dynamically receiving feedback associated with the interaction as the interaction occurs, wherein the feedback is either positive or negative with respect to the selection of the agent;
automatically updating the trained neural network based on the feedback as the feedback is received, wherein updating includes adjusting the one or more numeric weights associated with the selection of the agent based on the feedback, wherein the modification to the one or more numeric weights increases or decreases the one or more numeric weights based on whether the feedback is positive or negative, and wherein updating the trained neural network improves how the trained neural network performs further agent selections; and
selecting a second agent using the updated neural network.