IP Library Granted Patent US 12669992
Granted Patent B2
US 12669992 · App. 18/534,238 · Granted Jun 30, 2026

System and method for determining potential issues with a software deployment

Inventors: Gary B. Duma (Royal Oak, MI); Sheel Saket (Warrenville, IL)
Assignee: Fidelity Information Services, LLC
G06F8/65G06F11/1433G06N20/00
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12669992
App. No.
18/534,238
Granted
Jun 30, 2026
Kind
B2
Abstract

Systems, devices, methods, and computer readable media for determining potential issues with a software upgrade are disclosed. In one implementation, the disclosed system includes at least one processor and at least one non-transitory memory storing instructions to perform operations when executed by the at least one processor. The operations include training a machine learning model on historic trouble ticket data; receiving data relating to a planned software upgrade; providing the received data to the trained machine learning model, wherein the trained machine learning model generates a tag value; calculating a risk score based on the tag value; and determining a client-specific action to be performed relating to the planned software upgrade and based on the risk score.

Claims (74)

1 . A system for determining potential issues with a software upgrade, comprising:

at least one processor; and

at least one non-transitory memory storing instructions to perform operations when executed by the at least one processor including:

training a machine learning model on historic trouble ticket data, wherein the historic trouble ticket data includes issues raised in connection with a prior software upgrade, how the raised issues were resolved, and client sentiment regarding the issues raised and their resolution;

receiving data relating to a planned software upgrade;

providing the received data to the trained machine learning model, wherein:

the trained machine learning model generates a tag value;

the tag value is associated with a trouble ticket; and

the tag value is used to assign the trouble ticket to a class of potential client impact categories based on historical trouble ticket patterns;

calculating a risk score for the trouble ticket based on information in the trouble ticket and the associated tag value, wherein calculating the risk score includes:

determining a likelihood that the planned software upgrade will cause client-reported problems based on similarity to historical software upgrade patterns; and

correlating the planned software upgrade with historical client-reported issues; and

determining an action plan to be performed relating to the planned software upgrade and based on the risk score, wherein the action plan includes generating client-specific communications based on historical client sentiment analysis derived from historical trouble ticket resolutions.

2 . The system of claim 1 , wherein the historic trouble ticket data includes software update data, bug tracking data, source code host data, and project management data.

3 . The system of claim 1 , wherein the received data includes software update data, source code host data, and project management data.

4 . The system of claim 1 , wherein providing the received data to the trained machine learning model includes performing one or more of: data ingestion; extraction, transformation, and loading; natural language processing; and time matching.

5 . The system of claim 4 , wherein performing time matching includes matching a time of an historic software upgrade with a time of bug tracking data and a time of project management data.

6 . The system of claim 1 , wherein the instructions further include:

classifying potential issues with the planned software upgrade based on the tag value.

7 . The system of claim 6 , wherein the risk score is based on a classification and the tag value.

8 . The system of claim 1 , wherein the instructions further include:

comparing the risk score to a first threshold; and

on a condition that the risk score exceeds the first threshold, determining the action plan to be performed relating to the planned software upgrade and based on the risk score.

9 . The system of claim 8 , wherein the instructions further include:

calculating a sentiment score based on prior client communications;

on a condition that the sentiment score exceeds a second threshold and is less than a third threshold, generating a client communication by a trained second machine learning model regarding the planned software upgrade; and

on a condition that the sentiment score exceeds the second threshold and the third threshold, generating and sending an alert to a client support team to contact the client.

10 . The system of claim 1 , wherein the action plan is generated by a second machine learning model, the second machine learning model having been trained on historical communications with the client.

11 . A computer-implemented method for determining potential issues with a software upgrade, comprising:

training a machine learning model on historic trouble ticket data, wherein the historic trouble ticket data includes issues raised in connection with a prior software upgrade, how the raised issues were resolved, and client sentiment regarding the issues raised and their resolution;

receiving data relating to a planned software upgrade;

providing the received data to the trained machine learning model, wherein:

the trained machine learning model generates a tag value;

the tag value is associated with a trouble ticket; and

the tag value is used to assign the trouble ticket to a class of potential client impact categories based on historical trouble ticket patterns;

calculating a risk score for the trouble ticket based on information in the trouble ticket and the associated tag value, wherein calculating the risk score includes:

determining a likelihood that the planned software upgrade will cause client-reported problems based on similarity to historical software upgrade patterns; and

correlating the planned software upgrade with historical client-reported issues; and

determining an action plan to be performed relating to the planned software upgrade and based on the risk score, wherein the action plan includes generating client-specific communications based on historical client sentiment analysis derived from historical trouble ticket resolutions.

12 . The method of claim 11 , wherein the historic trouble ticket data includes software update data, bug tracking data, source code host data, and project management data.

13 . The method of claim 11 , wherein the received data includes software update data, source code host data, and project management data.

14 . The method of claim 11 , wherein providing the received data to the trained machine learning model includes performing one or more of: data ingestion; extraction, transformation, and loading; natural language processing; and time matching.

15 . The method of claim 14 , wherein performing time matching includes matching a time of an historic software upgrade with a time of bug tracking data and a time of project management data.

16 . The method of claim 11 , further comprising:

classifying potential issues with the planned software upgrade based on the tag value.

17 . The method of claim 16 , wherein the risk score is based on a classification and the tag value.

18 . The method of claim 11 , further comprising:

comparing the risk score to a first threshold; and

on a condition that the risk score exceeds the first threshold, determining the action plan to be performed relating to the planned software upgrade and based on the risk score.

19 . The method of claim 18 , further comprising:

calculating a sentiment score based on prior client communications;

on a condition that the sentiment score exceeds a second threshold and is less than a third threshold, generating a client communication by a trained second machine learning model regarding the planned software upgrade; and

on a condition that the sentiment score exceeds the second threshold and the third threshold, generating and sending an alert to a client support team to contact the client.

20 . A non-transitory computer-readable storage medium comprising instructions for determining potential issues with a software upgrade, wherein when executed by a processor perform operations comprising:

training a machine learning model on historic trouble ticket data, wherein the historic trouble ticket data includes issues raised in connection with a prior software upgrade, how the raised issues were resolved, and client sentiment regarding the issues raised and their resolution;

receiving data relating to a planned software upgrade;

providing the received data to the trained machine learning model, wherein:

the trained machine learning model generates a tag value;

the tag value is associated with a trouble ticket; and

the tag value is used to assign the trouble ticket to a class of potential client impact categories based on historical trouble ticket patterns;

calculating a risk score for the trouble ticket based on information in the trouble ticket and the associated tag value, wherein calculating the risk score includes:

determining a likelihood that the planned software upgrade will cause client-reported problems based on similarity to historical software upgrade patterns; and

correlating the planned software upgrade with historical client-reported issues; and

determining an action plan to be performed relating to the planned software upgrade and based on the risk score, wherein the action plan includes generating client-specific communications based on historical client sentiment analysis derived from historical trouble ticket resolutions.

21 . The non-transitory computer-readable storage medium of claim 20 , wherein providing the received data to the trained machine learning model includes performing one or more of: data ingestion; extraction, transformation, and loading;

natural language processing; and time matching.

22 . The non-transitory computer-readable storage medium of claim 21 , wherein performing time matching includes matching a time of an historic software upgrade with a time of bug tracking data and a time of project management data.

23 . The non-transitory computer-readable storage medium of claim 20 , wherein the operations further comprise:

comparing the risk score to a first threshold; and

on a condition that the risk score exceeds the first threshold, determining the action plan to be performed relating to the planned software upgrade and based on the risk score.

24 . The non-transitory computer-readable storage medium of claim 23 , wherein the operations further comprise:

calculating a sentiment score based on prior client communications;

on a condition that the sentiment score exceeds a second threshold and is less than a third threshold, generating a client communication by a trained second machine learning model regarding the planned software upgrade; and

on a condition that the sentiment score exceeds the second threshold and the third threshold, generating and sending an alert to a client support team to contact the client.