IP Library › Granted Patent US 12,536,236
Granted Patent B2
US 12,536,236 · App. 19/015,318 · Granted Jan 27, 2026

Systems and methods for generating document formatting and content using generative AI

Inventors: Andre Frade (London, GB); Marcus Kaiser (London, GB); Maxime Labonne (London, GB); Sean Moran (London, GB); Jack Donnelly (Rye, NY); Daniel Roose (London, GB); Babak Khorshidi (Surrey, GB); Lily Mcinerney (Hoboken, NJ); Jacob Marold (Pound Ridge, NY); Harshavardhan Vemulapati (New York, NY)
Assignee: JPMORGAN CHASE BANK, N.A.
G06F16/93
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 12,536,236
App. No.
19/015,318
Granted
Jan 27, 2026
Kind
B2
Abstract

Systems and methods for providing an archive of precedents are provided. The archive of precedents includes determining precedent data; parsing a new-event XML file for string data, wherein the string data is related to a new event; generating a vector embedding from the string data; determining a similar embedding in a vector space, wherein the similar embedding is a vector embedding of an instance of precedent data; parsing XML files of the precedent and new event and using parsed data to generate an old-to-new value map; retrieving the corresponding XML file and the corresponding precedent-event document associated with the vector embedding of the instance of precedent data; and providing, as input to a natural language processing model, the list of text strings recovered from the editable form of the document of the precedent event and the old-to-new value map.

Claims (44)

1 . A method comprising:

providing, by a database manager executed by one or more processors, an archive of precedents, wherein the archive of precedents includes, for each of a plurality of events, precedent data, and wherein the precedent data for each of the plurality of events includes a corresponding open-format file and a corresponding precedent-event document, and each corresponding to a distinct representation of a precedent event;

parsing, by an overwriting application executed by one or more processors, a new-event open-format file for string data, wherein the string data is related to a new event;

generating, by the overwriting application, a vector embedding from the string data;

determining, by the overwriting application, a similar embedding in a vector space using a cosine distance wherein the similar embedding is a vector embedding or representation of an instance of precedent data;

parsing, by the overwriting application, open-format files of the precedent and new event and using parsed data to generate an old-to-new value map;

retrieving, by the overwriting application, the corresponding open-format file and the corresponding precedent-event document(s) associated with the vector embedding of the instance of precedent data;

providing, by the overwriting application, as input to a natural language processing model, the list of text strings recovered from the editable form of the document of the precedent event and the old-to-new value map;

modifying, by the natural language processing model executed by one or more processors, the corresponding open-format file and the corresponding precedent-event document by including element replacement values according to the old-to-new value maps, and

receiving, by a user interface through the overwriting application and from the natural language processing model, a modified version of the corresponding open-format file and a modified version of the corresponding precedent-event document.

2 . The method of claim 1 , comprising:

saving the modified version of the corresponding open-format file and the modified version of the corresponding precedent-event document as a new open-format file and a new event document, respectively.

3 . The method of claim 2 , comprising:

adding the new open-format file and the new event document as precedent data to the archive of precedents.

4 . The method of claim 1 , wherein the element replacement values include a text element update.

5 . The method of claim 1 , wherein the element replacement values include a table element update.

6 . The method of claim 1 , wherein the element replacement values include an illustration element update.

7 . A system comprising at least one computer including a processor and a memory, wherein the at least one computer is configured to:

provide, by a database manager executed by one or more processors, an archive of precedents, wherein the archive of precedents includes, for each of a plurality of events, precedent data, and wherein the precedent data for each of the plurality of events includes a corresponding open-format file and a corresponding precedent-event document, and each corresponding to a distinct representation of a precedent event;

parse, by an overwriting application executed by one or more processors, a new-event open-format file for string data, wherein the string data is related to a new event;

generate, by the overwriting application, a vector embedding from the string data;

determine, by the overwriting application, a similar embedding in a vector space using a cosine distance wherein the similar embedding is a vector embedding or representation of an instance of precedent data;

parse, by the overwriting application, open-format files of the precedent and new event and using parsed data to generate an old-to-new value map;

retrieve, by the overwriting application, the corresponding open-format file and the corresponding precedent-event document(s) associated with the vector embedding of the instance of precedent data;

provide, by the overwriting application, as input to a natural language processing model, the list of text strings recovered from the editable form of the document of the precedent event and the old-to-new value map;

modify, by the natural language processing model executed by one or more processors, the corresponding open-format file and the corresponding precedent-event document by including element replacement values according to the old-to-new value maps, and

receive, by a user interface through the overwriting application and from the natural language processing model, a modified version of the corresponding open-format file and a modified version of the corresponding precedent-event document.

8 . The system of claim 7 , wherein the at least one computer is configured to:

save the modified version of the corresponding open-format file and the modified version of the corresponding precedent-event document as a new open-format file and a new event document, respectively.

9 . The system of claim 8 , wherein the at least one computer is configured to:

add the new open-format file and the new event document as precedent data to the archive of precedents.

10 . The system of claim 7 , wherein the element replacement values include a text element update.

11 . The system of claim 7 , wherein the element replacement values include a table element update.

12 . The system of claim 7 , wherein the element replacement values include an illustration element update.

13 . A method comprising:

providing, by a database manager executed by one or more processors, an archive of precedents, wherein the archive of precedents includes, for each of a plurality of events, precedent data, and wherein the precedent data for each of the plurality of events includes a corresponding open-format file and a corresponding precedent-event document, and each corresponding to a distinct representation of a precedent event;

parsing, by an overwriting application executed by one or more processors, a new-event open-format file for string data, wherein the string data is related to a new event;

generating, by the overwriting application, a vector embedding from the string data;

determining, by the overwriting application, a similar embedding in a vector space using an output of a cross encoder model, wherein the similar embedding is a vector embedding or representation of an instance of precedent data;

parsing, by the overwriting application, open-format files of the precedent and new event and using parsed data to generate an old-to-new value map;

retrieving, by the overwriting application, the corresponding open-format file and the corresponding precedent-event document(s) associated with the vector embedding of the instance of precedent data;

providing, by the overwriting application, as input to a natural language processing model, the list of text strings recovered from the editable form of the document of the precedent event and the old-to-new value map;

modifying, by the natural language processing model executed by one or more processors, the corresponding open-format file and the corresponding precedent-event document by including element replacement values according to the old-to-new value maps, and

receiving, by a user interface through the overwriting application and from the natural language processing model, a modified version of the corresponding open-format file and a modified version of the corresponding precedent-event document.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Oct 10, 2025
From: FRADE, ANDRE; KAISER, MARCUS; LABONNE, MAXIME; MORAN, SEAN; DONNELLY, JACK; ROOSE, DANIEL; KHORSHIDI, BABAK; MCINERNEY, LILY; MAROLD, JACOB; VEMULAPATI, HARSHAVARDHAN
To: JPMORGAN CHASE BANK, N.A.
Reel/Frame 072534/0310 →
Continuity (2)
Provisional Application 63619177 · Jan 9, 2024
Related Publication 20250225188A1 · Jul 10, 2025
References Cited (4)
US 12417214B1 · Zhang · 2025 [cited by examiner]
US 20040268246A1 · Leban · 2004 [cited by examiner]
US 20230315993A1 · Nieborowski · 2023 [cited by examiner]
CN 112380827B · 2022 [cited by examiner]