IP Library Granted Patent US 12682164
Granted Patent B2
US 12682164 · App. 18/192,699 · Granted Jul 14, 2026

Chat support platform having automatic keyword correction

Inventor: Barath Jayaraman (Charlotte, NC)
Assignee: Truist Bank
G06F40/232G06F40/166G06F40/268G06F40/279G06Q30/0281
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 12682164
App. No.
18/192,699
Granted
Jul 14, 2026
Kind
B2
Abstract

A computing device, a computer program product, and a computer-implemented method for delivering enhanced financial services and, more particularly, for facilitating a virtual chat communication session between a user and a virtual support agent which automatically identifies and corrects keywords which the user may misspelled. A pre-stored list of valid keywords is maintained and is used to identify which keywords are misspelled. Misspelled keywords may be automatically corrected.

Claims (55)

1 . A server computing system, comprising:

one or more processors; and

a non-transitory memory coupled to the one or more processors, the non-transitory memory including a set of instructions of computer-executable program code, which when executed by the one or more processors, causes the server computing system to:

initiate a dedicated process by displaying a chat interface on a user interface of an authenticated client device executing an enterprise mobile application or enterprise desktop application to facilitate a virtual chat communication session between a user and a chatbot;

receive, from the authenticated client device via the chat interface, a user command;

parse, using a machine learning (ML) module executing a trained ML model and contemporaneously with the virtual chat communication session, the user command to determine constituent words of the user command;

process, using the trained ML model contemporaneously with the virtual chat communication session, the constituent words to determine a set of keywords from the constituent words;

determine, using the trained ML model contemporaneously with the virtual chat communication session, a corresponding valid keyword from a plurality of pre-stored keywords for each of the set of keywords while disregarding other portions of the constituent words;

detect, using the trained ML model contemporaneously with the virtual chat communication session, a misspelled keyword from the set of keywords based on a comparison with the corresponding valid keyword;

determine, using the ML module executing the trained ML model contemporaneously with the virtual chat communication session in response to detecting the misspelled keyword, a candidate keyword out of the plurality of pre-stored keywords which is a closest match out of the plurality of pre-stored keywords to the misspelled keyword;

determine, contemporaneously with the virtual chat communication session, whether at least a predetermined number of letters in the misspelled keyword match the predetermined number of letters in the candidate keyword;

replace, responsive to determining that at least the predetermined number of letters in the misspelled keyword match the predetermined number of letters in the candidate keyword and contemporaneously with the virtual chat communication session, the misspelled keyword in the user command with the candidate keyword to generate a corrected command; and

cause, responsive to determining that at least the predetermined number of letters in the misspelled keyword match the predetermined number of letters in the candidate keyword and contemporaneously with the virtual chat communication session, implementation of the corrected command.

2 . The server computing system of claim 1 , wherein the predetermined number of letters is at least 80% of a number of letters in the misspelled keyword.

3 . The server computing system of claim 1 , wherein the trained ML model is configured to:

process the constituent words to determine personalized information words from the constituent words, wherein the personalized information words are disregarded when the trained ML model determines corresponding valid keywords for each of the set of keywords.

4 . The server computing system of claim 1 , wherein the set of instructions, which when executed by the one or more processors, causes a display of the candidate keyword on the chat interface.

5 . The server computing system of claim 1 , wherein the set of instructions, which when executed by the one or more processors, enables the user to enter a configuration command in the chat interface to set the predetermined number of letters.

6 . The server computing system of claim 1 , wherein the set of instructions, which when executed by the one or more processors, causes the server computing system to select, responsive to determining that more than one keyword out of the plurality of pre-stored keywords qualify as a closest match to the misspelled keyword, a particular keyword out of the plurality of pre-stored keywords that is used with a highest frequency as the candidate keyword.

7 . The server computing system of claim 1 , wherein the set of instructions, which when executed by the one or more processors, causes the server computing system to set the predetermined number of letters as a default setting for further virtual chat communication sessions.

8 . A computer program product comprising at least one non-transitory computer readable medium having a set of instructions of computer-executable program code, which when executed by one or more processors of a server computing system, causes the server computing system to:

initiate a dedicated process by displaying a chat interface on a user interface of an authenticated client device executing an enterprise mobile application or enterprise desktop application to facilitate a virtual chat communication session between a user and a chatbot;

receive, from the authenticated client device via the chat interface, a user command;

parse, using a machine learning (ML) module executing a trained ML model and contemporaneously with the virtual chat communication session, the user command to determine constituent words of the user command;

process, using the trained ML model contemporaneously with the virtual chat communication session, the constituent words to determine a set of keywords from the constituent words;

determine, using the trained ML model contemporaneously with the virtual chat communication session, a corresponding valid keyword from a plurality of pre-stored keywords for each of the set of keywords while disregarding other portions of the constituent words;

detect, using the trained ML model contemporaneously with the virtual chat communication session, a misspelled keyword from the set of keywords based on a comparison with the corresponding valid keyword;

determine, using the ML module executing the trained ML model contemporaneously with the virtual chat communication session in response to detecting the misspelled keyword, a candidate keyword out of the plurality of pre-stored keywords which is a closest match out of the plurality of pre-stored keywords to the misspelled keyword;

determine, contemporaneously with the virtual chat communication session, whether at least a predetermined number of letters in the misspelled keyword match the predetermined number of letters in the candidate keyword;

replace, responsive to determining that at least the predetermined number of letters in the misspelled keyword match the predetermined number of letters in the candidate keyword and contemporaneously with the virtual chat communication session, the misspelled keyword in the user command with the candidate keyword to generate a corrected command; and

cause, responsive to determining that at least the predetermined number of letters in the misspelled keyword match the predetermined number of letters in the candidate keyword and contemporaneously with the virtual chat communication session, implementation of the corrected command.

9 . The computer program product of claim 8 , wherein the predetermined number of letters is at least 80% of a number of letters in the misspelled keyword.

10 . The computer program product of claim 8 , wherein the trained ML model is configured to:

process the constituent words to determine personalized information words from the constituent words, wherein the personalized information words are disregarded when the trained ML model determines corresponding valid keywords for each of the set of keywords.

11 . The computer program product of claim 8 , wherein the set of instructions, which when executed by the one or more processors, causes a display of the candidate keyword on the chat interface.

12 . The computer program product of claim 8 , wherein the set of instructions, which when executed by the one or more processors, enables the user to enter a configuration command in the chat interface to set the predetermined number of letters.

13 . The computer program product of claim 8 , wherein the set of instructions, which when executed by the one or more processors, causes the server computing system to select, responsive to determining that more than one keyword out of the plurality of pre-stored keywords qualify as a closest match to the misspelled keyword, a particular keyword out of the plurality of pre-stored keywords that is used with a highest frequency as the candidate keyword.

14 . The computer program product of claim 8 , wherein the set of instructions, which when executed by the one or more processors, causes the server computing system to set the predetermined number of letters as a default setting for further virtual chat communication sessions.

15 . A computer implemented method for implementation by a server computing system, the computer-implemented method comprising:

initiating a dedicated process by displaying a chat interface on a user interface of an authenticated client device executing an enterprise mobile application or enterprise desktop application to facilitate a virtual chat communication session between a user and a chatbot;

receiving, from the authenticated client device via the chat interface, a user command;

parsing, using a machine learning (ML) module executing a trained ML model and contemporaneously with the virtual chat communication session, the user command to determine constituent words of the user command;

processing, using the trained ML model contemporaneously with the virtual chat communication session, the constituent words to determine a set of keywords from the constituent words;

determining, using the trained ML model contemporaneously with the virtual chat communication session, a corresponding valid keyword from a plurality of pre-stored keywords for each of the set of keywords while disregarding other portions of the constituent words;

detecting, using the trained ML model contemporaneously with the virtual chat communication session, a misspelled keyword from the set of keywords based on a comparison with the corresponding valid keyword;

determining, using the ML module executing the trained ML model contemporaneously with the virtual chat communication session in response to detecting the misspelled keyword, a candidate keyword out of the plurality of pre-stored keywords which is a closest match out of the plurality of pre-stored keywords to the misspelled keyword;

determining, contemporaneously with the virtual chat communication session, whether at least a predetermined number of letters in the misspelled keyword match the predetermined number of letters in the candidate keyword;

replacing, responsive to determining that at least the predetermined number of letters in the misspelled keyword match the predetermined number of letters in the candidate keyword and contemporaneously with the virtual chat communication session, the misspelled keyword in the user command with the candidate keyword to generate a corrected command; and

causing, responsive to determining that at least the predetermined number of letters in the misspelled keyword match the predetermined number of letters in the candidate keyword and contemporaneously with the virtual chat communication session, implementation of the corrected command.

16 . The computer implemented method of claim 15 , wherein the predetermined number of letters is at least 80% of a number of letters in the misspelled keyword.

17 . The computer implemented method of claim 15 , wherein the trained ML model is configured to:

process the constituent words to determine a set of keywords and personalized information words from the constituent words, wherein the personalized information words are disregarded when the trained ML model determines corresponding valid keywords for each of the set of keywords.

18 . The computer implemented method of claim 15 , further comprising causing a display of the candidate keyword on the chat interface.

19 . The computer implemented method of claim 15 , further comprising enabling the user to enter a configuration command in the chat interface to set the predetermined number of letters.

20 . The computer implemented method of claim 15 , further comprising, responsive to determining that more than one keyword out of the plurality of pre-stored keywords qualify as a closest match to the misspelled keyword, selecting a particular keyword out of the plurality of pre-stored keywords that is used with a highest frequency as the candidate keyword.