Granted Patent B1

US 12,591,938 · App. 18/901,375 · Granted Mar 31, 2026

Inventors: Nikhil Dinesh Yajaman (Denton, TX); Morgan J. Finley (St Louis Park, MN); Micah Mclean (Gilbert, AZ); Anirban Chowdhury (Frisco, TX); Mantu Kumar (Jamshedpur, IN)

Assignee: Express Scripts Strategic Development, Inc.

G06Q40/082G06F16/2358G06Q30/0637G06Q40/08G06Q40/09G16H20/10

Abstract

Claims (120)

1 . A method comprising:

receiving an indication that a user record associated with a user has been updated with a first change;

determining a confidence score that the first change is valid, wherein the confidence score does not meet a confidence threshold;

filtering a set of communications associated with the user to create a subset of communications, wherein filtering includes determining whether a respective communication of the set of communications includes a communication topic of the first change;

determining, via an analysis of the subset of communications by a first machine learning model, whether the first change has met one or more of a set of validity criteria, wherein:

a communication of the subset of communications includes a first portion, a second portion, and a third portion, and

the first portion and the third portion are analyzed by the first machine learning model before the second portion is analyzed; and

in response to a determination that the first change has not met the set of validity criteria:

based on a first outcome of a set of reporting criteria, automatically generating a prompt with a first suggested revision of the first change; and

based on a second outcome of the set of reporting criteria, automatically revising the first change with the first suggested revision.

2 . The method of claim 1 further comprising, based on the first outcome of the set of reporting criteria:

receiving a verification input corresponding to the prompt;

in response to a modification input corresponding to the prompt, revising the first suggested revision;

in response to an approval input corresponding to the prompt, accepting the first suggested revision; and

in response to a rejection input corresponding to the prompt, rejecting the first suggested revision.

3 . The method of claim 1 wherein:

the confidence score is determined by a second machine learning model;

the confidence score is based on a set of historical data associated with the user, including:

a current time of year,

a level of novelty associated with the first change,

a set of previous values associated with the user record,

a set of dates associated with the set of previous values, and

a quantity of deliveries associated with the set of previous values; and

the method further comprises, updating a set of training data for the second machine learning model based on at least one of:

the confidence score,

a user input corresponding to the prompt, or

the first change.

4 . The method of claim 1 wherein the filtering is performed by a third machine learning model.

5 . The method of claim 1 further comprising, in response to a determination that the first change is valid:

generating an indication that the first change has been verified, and

automatically generating and transmitting a report associated with the first change.

6 . The method of claim 1 further comprising, in response to a determination that the confidence score has met a confidence threshold:

generating an indication that the first change has been verified, and

automatically generating and transmitting a report associated with the first change.

7 . The method of claim 1 further comprising digitizing the subset of communications, wherein digitizing the subset of communications includes at least one of:

transcribing a respective communication,

performing an analysis via machine vision on the respective communication,

translating the communication,

processing text, audio, video, or image files, or

performing optical character recognition.

8 . The method of claim 1 wherein the set of reporting criteria includes:

a criterion that is met when a manufacturer associated with a product or service consumed by the user requires reports associated with the product or service to be transmitted to the manufacturer,

a criterion that is met when user data must be changed by an authorized agent, and

a criterion that is met when a user is associated with a product or service that meets a set of restriction requirements.

9 . The method of claim 1 wherein the set of validity criteria includes:

a criterion that is met when a user requests the first change in a respective communication of the subset of communications, and

a criterion that is met when the first change matches a set of data in the respective communication.

10 . The method of claim 1 further comprising, in response to a determination that the first change has met the set of validity criteria, automatically preparing a physical product delivery.

11 . The method of claim 1 wherein:

the first portion includes a beginning portion of the communication of the subset of communications,

the second portion includes a middle portion of the communication of the subset of communications, and

the third portion includes an end portion of the communication of the subset of communications.

12 . A non-transitory computer-readable medium storing processor-executable instructions, wherein the instructions include:

receiving an indication that a user record associated with a user has been updated with a first change;

determining a confidence score that the first change is valid; and

in response to a determination that the confidence score has not met a confidence threshold:

filtering a set of communications associated with the user to create a subset of communications, wherein filtering includes determining whether a respective communication of the set of communications includes a communication topic of the first change;

determining, via an analysis of the subset of communications by a first machine learning model, whether the first change has met one or more of a set of validity criteria, wherein:

a communication of the subset of communications includes a first portion, a second portion, and a third portion, and

the first portion and the third portion are analyzed by the first machine learning model before the second portion is analyzed; and

in response to a determination that the first change has not met the set of validity criteria:

based on a first outcome of a set of reporting criteria, automatically generating a prompt with a first suggested revision of the first change; and

based on a second outcome of the set of reporting criteria, automatically revising the first change with the first suggested revision.

13 . The non-transitory computer-readable medium of claim 12 wherein the instructions include, based on the first outcome of the set of reporting criteria:

receiving a verification input corresponding to the prompt;

in response to a modification input corresponding to the prompt, revising the first suggested revision;

in response to an approval input corresponding to the prompt, accepting the first suggested revision; and

in response to a rejection input corresponding to the prompt, rejecting the first suggested revision.

14 . The non-transitory computer-readable medium of claim 12 wherein:

the confidence score is determined by a second machine learning model;

the confidence score is based on a set of historical data associated with the user, including:

a current time of year,

a level of novelty associated with the first change,

a set of previous values associated with the user record,

a set of dates associated with the set of previous values, and

a quantity of deliveries associated with the set of previous values; and

the instructions include, updating a set of training data for the second machine learning model based on at least one of:

the confidence score,

a user input corresponding to the prompt, or

the first change.

15 . The non-transitory computer-readable medium of claim 12 wherein the instructions include, in response to a determination that the first change is valid:

generating an indication that the first change has been verified, and

automatically generating and transmitting a report associated with the first change.

16 . The non-transitory computer-readable medium of claim 12 wherein the set of reporting criteria includes:

a criterion that is met when a manufacturer associated with a product or service consumed by the user requires reports associated with the product or service to be transmitted to the manufacturer,

a criterion that is met when user data must be changed by an authorized agent, and

a criterion that is met when a user is associated with a product or service that meets a set of restriction requirements.

17 . A system comprising:

memory hardware configured to store instructions; and

processor hardware configured to execute instructions stored by the memory hardware, wherein the instructions include:

receiving an indication that a user record associated with a user has been updated with a first change;

determining a confidence score that the first change is valid; and

in response to a determination that the confidence score has not met a confidence threshold:

filtering a set of communications associated with the user to create a subset of communications, wherein filtering includes determining whether a respective communication of the set of communications includes a communication topic of the first change;

determining, via an analysis of the subset of communications by a first machine learning model, whether the first change has met one or more of a set of validity criteria, wherein:

a communication of the subset of communications includes a first portion, a second portion, and a third portion, and

the first portion and the third portion are analyzed by the first machine learning model before the second portion is analyzed; and

in response to a determination that the first change has not met the set of validity criteria:

based on a first outcome of a set of reporting criteria, automatically generating a prompt with a first suggested revision of the first change; and

based on a second outcome of the set of reporting criteria, automatically revising the first change with the first suggested revision.

18 . The system of claim 17 wherein the instructions include, based on the first outcome of the set of reporting criteria:

receiving a verification input corresponding to the prompt;

in response to a modification input corresponding to the prompt, revising the first suggested revision;

in response to an approval input corresponding to the prompt, accepting the first suggested revision; and

in response to a rejection input corresponding to the prompt, rejecting the first suggested revision.

19 . The system of claim 17 wherein:

the confidence score is determined by a second machine learning model;

the confidence score is based on a set of historical data associated with the user, including:

a current time of year,

a level of novelty associated with the first change,

a set of previous values associated with the user record,

a set of dates associated with the set of previous values, and

a quantity of deliveries associated with the set of previous values; and

the instructions include, updating a set of training data for the second machine learning model based on at least one of:

the confidence score,

a user input corresponding to the prompt, or

the first change.

20 . The system of claim 17 wherein the instructions include, in response to a determination that the first change is valid:

generating an indication that the first change has been verified, and

automatically generating and transmitting a report associated with the first change.

References Cited (19)