System and method of navigating a digital platform with a customized virtual assistant
View Patent ↗A method and system of navigating a digital platform with a customized virtual assistant is provided. More specifically, a firm, with a digital platform such as a website, an app, a self-hosted app, and/or an API toolkit will be able to provide its users with a method to ask, by voice on an app/website, type on an app/website, and/or speak over an Interactive Voice Response (IVR) call, to access these entities. Further, the method may include prompting to enter at least one user query through a digital user interface. Further, the method may include retrieving a query response using an assistance engine, based on the user query. Further, the retrieving may be based on a contextually-aware Large Language Model. Further, the method may include outputting, using the assistance engine, and the query response to the digital platform.
1 . A method of navigating a digital platform with a customized virtual assistant, the method comprising the steps of:
(A) providing a digital user interface, an assistance engine, and digital platform information managed by a computing system, wherein the assistance engine is operatively implemented with a contextually-aware large language model, and wherein the contextually-aware large language model is operatively conditioned with at least one knowledge-basing process, and wherein the knowledge-basing process is configured to source external information into the contextually-aware large language model, and wherein the digital user interface is a digital platform configured to be virtually navigated by a user, and wherein the digital platform information is any information accessible through the digital user interface, and wherein the digital user interface is accessible via a user personal computing (PC) device, and wherein the digital platform is a website with a plurality of navigable webpages, and wherein the website is functionally independent from the contextually-aware large language model;
(B) prompting to activate the assistance engine through the digital user interface by displaying a virtual microphone icon, a virtual phone keypad, a virtual chat icon, or a combination thereof with the user PC device;
(C) prompting to enter at least one user query through the digital user interface, if the assistance engine is selected to be activated through the digital user interface, wherein the user query is at least one voice message and/or at least one textual message;
(D) retrieving a query response to the user query with the contextually-aware large language model by referencing the digital platform information with the knowledge-basing process, if the user query is entered through the digital user interface, wherein the contextually-aware large language model and the digital platform information are functionally independent from each other, and wherein the query response is mapped through the digital platform information by the contextually-aware large language model, and wherein the query response includes a webpage identifier and/or a webpage action;
(E) outputting the query response with the digital user interface;
sequentially executing steps (B) through (E);
hosting a specific webpage associated with the webpage identifier through the user PC device after step (E), if the query response includes the webpage identifier; and
executing the webpage action for a specific webpage with the user PC device after step (E), if the query response includes the webpage action.
2 . The method as claimed in claim 1 comprising the steps of:
receiving the voice message with a microphone of the user PC device after step (C);
relaying the voice message from the user PC device to the computing system; and
processing the voice message as the user query with the computing system before step (D).
3 . The method as claimed in claim 1 comprising the steps of:
(F) receiving the voice message with a microphone of the user PC device after step (C);
(G) relaying the voice message from the user PC device to the computing system;
(H) generating at least one message follow-up with the computing system, if the message follow-up is triggered by the voice message;
(I) prompting to enter an audible response to the message follow-up through the digital user interface;
(J) relaying the audible response from the user PC device to the computing system, if the audible response is received through the microphone of the user PC device;
(K) executing a plurality of iterations of steps (F) through (J) as the audible response becomes the voice message in each iteration, until the voice message does not trigger the message follow-up; and
(L) processing the voice message as the user query with the computing system before step (D).
4 . The method as claimed in claim 1 comprising the steps of:
receiving the textual message through the digital user interface after step (C);
relaying the textual message from the user PC device to the computing system; and
processing the textual message as the user query with the computing system before step (D).
5 . The method as claimed in claim 1 , wherein the webpage identifier is selected from a group consisting of: a uniform resource locator (URL), a uniform resource identifier (URI), an application programming interface (API) endpoint, and an application webpage identifier.
6 . The method as claimed in claim 1 , wherein the webpage action is executed with at least one specific API.
7 . The method as claimed in claim 1 comprising the step of:
executing step (D) with the computing system by statically mapping the user query through the digital platform information.
8 . The method as claimed in claim 1 comprising the step of:
executing step (D) with the computing system by dynamically mapping the user query through the knowledge-basing process.
9 . The method as claimed in claim 1 , wherein the knowledge-basing process is a retrieval augmented generation process.
10 . The method as claimed in claim 1 , wherein the knowledge-basing process is a finetuning process.
11 . The method as claimed in claim 1 comprising the steps of:
providing the computing system as the user PC device; and
managing the digital user interface, the assistance engine, and the digital platform information with the user PC device during steps (A) through (E).
12 . The method as claimed in claim 1 comprising the steps of:
providing the computing system as a single server; and
managing the digital user interface, the assistance engine, and the digital platform information with the single server during steps (A) through (E).
13 . The method as claimed in claim 1 comprising the steps of:
providing the computing system as a plurality of servers, wherein the plurality of servers includes a primary server and a secondary server;
managing the digital platform information with the primary server during steps (A) through (E); and
managing the assistance engine with the secondary server during steps (A) through (E).
14 . The method as claimed in claim 1 comprising the steps of:
providing the computing system as a plurality of servers, wherein the plurality of servers includes a primary server and a secondary server;
managing the digital user interface with the primary server during steps (A) through (E); and
managing the assistance engine and the digital platform information with the secondary server during steps (A) through (E).
15 . The method as claimed in claim 1 comprising the steps of:
providing the computing system as a plurality of servers, wherein the plurality of servers includes a primary server, a secondary server, and a tertiary server;
managing the digital user interface with the primary server during steps (A) through (E);
managing the digital platform information with the secondary server during steps (A) through (E); and
managing the assistance engine with the tertiary server during steps (A) through (E).