Granted Patent B2

US 12,705,266 · App. 18/353,449 · Granted Aug 11, 2026

Inventors: Paul Sheward (Chicago, IL); Chung-Sheng Li (Scarsdale, NY); Scott Likens (Austin, TX); Saverio Fato (Wilton Manors, FL); Joseph Doyle Harrington (Granite Bay, CA); Joseph David Voyles (Louisville, KY); Jonathan B. Rhine (Suffern, NY); Alexander Nicholas Boldizsar (Woodstock, CT); Winnie Cheng (West New York, NJ); Todd Christopher Morrill (Newburgh, NY); Yuan Wan (Irvine, CA); William Spotswood Seward (Los Angeles, CA)

Assignee: PwC Product Sales LLC

G06F16/3329

Abstract

Claims (64)

1 . A method for providing a hybrid AI and human electronic communication interface, comprising:

receiving a first electronic transmission comprising a user query from a user device;

automatically generating, by processing the user query by a set of AI models, an automatic response to the user query;

electronically transmitting the automatic response to the user device;

receiving, from the user device, a second electronic transmission comprising a user input in response to the automatic response;

determining, based at least in part on processing the user input, that a set of criteria is met,

wherein determining that the set of criteria is met comprises evaluating a plurality of criteria,

wherein the plurality of criteria are evaluated based on a predetermined order, and

determining that the set of criteria is met when at least a threshold number of the plurality of criteria are satisfied,

wherein processing the user input comprises performing sentiment analysis on a text of the user input to determine a sentiment of the user input, and

wherein the set of criteria comprises the sentiment being a negative sentiment; and

in accordance with the determination that the set of criteria are met, automatically instantiating an electronic communication connection between the user device and a second user device.

2 . The method of claim 1 ,

wherein the user query is a first query,

wherein the user input comprises a second query,

wherein processing the user input comprises determining a similarity between the first query and the second query, and

wherein determining the similarity between the first query and the second query comprises:

generating a first text embedding for the first query and a second text embedding for the second query; and

determining a similarity score between the first text embedding and the second text embedding.

3 . The method of claim 1 , wherein the user input comprises one or more additional queries, wherein processing the user input comprises determining a quantity of the one or more additional queries, and wherein the set of criteria comprises the quantity satisfying a threshold number of queries.

4 . The method of claim 1 , wherein performing the sentiment analysis comprises detecting one or more keywords in the text of the user input.

5 . The method of claim 1 , wherein the user input comprises visual facial data of a user of the user device, wherein processing the user input comprises processing the visual facial data by a second set of AI models, and wherein the set of criteria comprises a negative sentiment output by the second set of AI models.

6 . The method of claim 1 , wherein the user input comprises negative feedback on the automatic response to the user query, and wherein the set of criteria comprises receiving the negative feedback.

7 . The method of claim 1 , further comprising storing the user query and an expert response to the user query provided by the second user device in a data store, wherein the user query is associated with the expert response in the data store.

8 . The method of claim 7 , further comprising automatically generating an improved automatic response to a subsequent user query based on the expert response stored in the data store.

9 . The method of claim 1 , wherein automatically generating the automatic response comprises determining that a similarity score between the user query and a stored query stored in a data store satisfies a similarity threshold, wherein the stored query is associated with a data structure comprising one or more supplemental queries of the stored query, and wherein the automatic response is generated based on information associated with the one or more supplemental queries.

10 . The method of claim 9 , further comprising:

receiving, from the second user device, subject matter information corresponding to an expert response provided to the user query by the second user device; and

updating the data structure with the subject matter information.

11 . The method of claim 10 , further comprising automatically generating an improved automatic response to a subsequent user query based on the updated data structure with the subject matter information.

12 . The method of claim 9 , wherein the data structure is a decision tree, wherein each of the one or more supplemental queries is a branching criteria for the decision tree, and wherein the information associated with the one or more supplemental queries are leaf nodes of the decision tree.

13 . The method of claim 1 , wherein automatically generating the automatic response comprises:

determining that a set of similarity scores between the user query and each of a plurality of stored queries stored in a data store is below a similarity threshold; and

generating one or more recommended queries to the user, wherein each of the one or more recommended queries is generated based on the user query.

14 . The method of claim 1 , wherein the second user device is selected from a set of user devices, wherein each user device of the set of user devices corresponds to one of a set of subject matter expert profiles, wherein the second user device is selected based on analyzing features in each of the set of subject matter expert profiles, wherein the features comprise at least one of:

an engagement history,

a success rate in generating responses to user queries, and

education and experience credentials.

15 . The method of claim 1 , further comprising:

transmitting, to the user device, a set of subject matter expert profiles to the user device; and

receiving, from the user device, a third electronic transmission comprising a selected subject matter expert profile, wherein the second user device corresponds to the selected subject matter expert profile.

16 . The method of claim 1 , further comprising:

automatically retrieving data on a user associated with the user device from a data store; and

wherein automatically instantiating the electronic communication comprises transmitting, to the second user device, the data on the user.

17 . A system for providing a hybrid AI and human electronic communication interface, the system comprising:

one or more processors; and

a memory coupled to the one or more processors comprising instructions executable by the one or more processors, the processors operable when executing the instructions to:

receive a first electronic transmission comprising a user query from a user device;

automatically generate, by processing the user query by a set of AI models, an automatic response to the user query;

electronically transmit the automatic response to the user device;

receive, from the user device, a second electronic transmission comprising a user input in response to the automatic response;

determine, based at least in part on processing the user input, that a set of criteria is met, wherein determining that the set of criteria is met comprises evaluating a plurality of criteria, wherein the plurality of criteria are evaluated based on a predetermined order, and determining that the set of criteria is met when at least a threshold number of the plurality of criteria are satisfied, and

wherein processing the user input comprises performing sentiment analysis on a text of the user input to determine a sentiment of the user input, and

wherein the set of criteria comprises the sentiment being a negative sentiment; and

in accordance with the determination that the set of criteria are met, automatically instantiate an electronic communication connection between the user device and a second user device.

18 . A non-transitory computer-readable storage medium storing instructions for providing a hybrid AI and human electronic communication interface, the instructions operable when executed by one or more processors of a system to cause the system to:

receive a first electronic transmission comprising a user query from a user device;

automatically generate, by processing the user query by a set of AI models, an automatic response to the user query;

electronically transmit the automatic response to the user device;

receive, from the user device, a second electronic transmission comprising a user input in response to the automatic response;

determine, based at least in part on processing the user input, that a set of criteria is met, wherein determining that the set of criteria is met comprises evaluating a plurality of criteria, wherein the plurality of criteria are evaluated based on a predetermined order, and determining that the set of criteria is met when at least a threshold number of the plurality of criteria are satisfied, and

wherein processing the user input comprises performing sentiment analysis on a text of the user input to determine a sentiment of the user input, and

wherein the set of criteria comprises the sentiment being a negative sentiment; and

in accordance with the determination that the set of criteria are met, automatically instantiate an electronic communication connection between the user device and a second user device.

Assignments (2)

Continuity (1)

References Cited (49)