User persona injection for task-oriented virtual assistants
An intelligent virtual assistant (IVA) is deployed to place a call/chat to customer service on behalf of the customer, thus saving them time and frustration. The IVA contacts a specific company over one or more channels, for example, chat, phone call, Application Programming Interface (API), or email, in order to complete open-ended task(s) requested by its user. Before the IVA contacts the company, a specific user profile is injected into an IVA dialog state. The IVA contacts the company or agency and answers customer service agent (CSA) questions by using the specific user profile provided for the call. The IVA then stores the task outcome for the user to review. If something prevents the task from succeeding, the IVA alerts the user that either it needs more information or the user may need to perform some action before the task can be completed, such as filling out or emailing a form.
1 . A method of automating data-based tasks on behalf of a user, comprising:
receiving task instructions comprising (i) a destination identifier that identifies a product or service provider and (ii) a channel identifier that identifies a communications channel associated with the product or service provider;
retrieving, from a database, one or more variables relevant to a task by querying the database based on an intent of the task instructions, wherein the one or more variables include a user identifier that is specific to the product or service provider;
loading an artificial intelligence (AI) model selected from one or more pre-trained models based on at least one of (i) the destination identifier, (ii) the channel identifier, and (iii) the intent of the task;
initiating a bi-directional communication with the product or service provider over the communications channel; and
generating a response to a query from the product or service provider by using the AI model to replace one or more variables in a response template with the one or more variables relevant to the task.
2 . The method of claim 1 , wherein the one or more variables relevant to the task include customer information associated with the user or dynamic information associated with the task.
3 . The method of claim 1 , wherein the communications channel is selected in accordance with a cost, speed, or urgency of the task.
4 . The method of claim 3 , wherein the communications channel is at least one of SMS messaging, email, posting on social channels, live chat, a live agent, or an API connection.
5 . The method of claim 1 , wherein the one or more pre-trained models are organized by specific task, entity, or industry.
6 . The method of claim 1 , further comprising authenticating the user with the product or service provider.
7 . The method of claim 1 , further comprising:
determining whether the response from the product or service provider requests unknown information, and if so, creating a new variable and obtaining new information to populate the new variable;
determining whether there are additional requests to complete the task, and if so, making the additional requests to the product or service provider; and
if there are no additional requests, notifying the user.
8 . The method of claim 1 , further comprising:
mapping a variable of the one or more variables relevant to the task to the one or more variables in the response template, wherein the one or more variables in the response template are delexicalized variable names corresponding to data within a user profile of the user.
9 . A computer system, comprising:
a memory comprising computer-executable instructions; and
a processor configured to execute the computer-executable instructions and cause the computer system to:
receive task instructions comprising (i) a destination identifier that identifies a product or service provider and (ii) a channel identifier that identifies a communications channel associated with the product or service provider;
retrieve, from a database, one or more variables relevant to a task by querying the database based on an intent of the task instructions, wherein the one or more variables include a user identifier that is specific to the product or service provider;
load an artificial intelligence (AI) model selected from one or more pre-trained models based on at least one of (i) the destination identifier, (ii) the channel identifier, and (iii) the intent of the task;
initiate a bi-directional communication with the product or service provider over the communications channel; and
generate a response to a query from the product or service provider by using the AI model to replace one or more variables in a response template with the one or more variables relevant to the task.
10 . The computer system of claim 9 , wherein the one or more variables relevant to the task include customer information associated with a user or dynamic information associated with the task.
11 . The computer system of claim 9 , wherein the communications channel is selected in accordance with a cost, speed, or urgency of the task.
12 . The computer system of claim 11 , wherein the communications channel is at least one of SMS messaging, email, posting on social channels, live chat, a live agent, or an API connection.
13 . The computer system of claim 9 , wherein the one or more pre- trained models
are organized by specific task, entity, or industry.
14 . The computer system of claim 9 , wherein the instructions further cause the computer system to authenticate a user with the product or service provider.
15 . The computer system of claim 9 , wherein the instructions further cause the computer system to:
determine whether the response from the product or service provider requests unknown information, and if so, creating a new variable and obtaining new information populate the new variable;
determine whether there are additional requests to complete the task, and if so, making the additional requests to the product or service provider; and
if there are no additional requests notifying a user.
16 . The computer system of claim 9 , wherein the instructions further cause the computer system to:
map a variable of the one or more variables relevant to the task to the one or more variables in the response template, wherein the one or more variables in the response template are delexicalized variable names corresponding to data within a user profile.
17 . A non-transitory computer readable medium comprising instructions that, when executed by a processor of a processing system, cause the processing system to:
receive task instructions comprising (i) a destination identifier that identifies a product or service provider and (ii) a channel identifier that identifies a communications channel associated with the product or service provider;
retrieve, from a database, one or more variables relevant to a task by querying the database based on an intent of the task instructions, wherein the one or more variables include a user identifier that is specific to the product or service provider;
load an artificial intelligence (AI) model selected from one or more pre-trained models based on at least one of (i) the destination identifier, (ii) the channel identifier, and (iii) the intent of the task;
initiate a bi-directional communication with the product or service provider over the communications channel; and
generate a response to a query from the product or service provider by using the AI model to replace one or more variables in a response template with the one or more variables relevant to the task.
18 . The non-transitory computer readable medium of claim 17 , wherein the one or more pre-trained models
are organized by specific task, entity, or industry.
19 . The non-transitory computer readable medium of claim 17 , wherein the instructions further cause the processing system to:
determine whether the response from the product or service provider requests unknown information, and if so, creating a new variable and obtaining new information populate the new variable;
determine whether there are additional requests to complete the task, and if so, making the additional requests to the product or service provider; and
if there are no additional requests, notifying a user.
20 . The non-transitory computer readable medium of claim 17 , wherein the instructions further cause the processing system to:
map a variable of the one or more variables relevant to the task to the one or more variables in the response template, wherein the one or more variables in the response template are delexicalized variable names corresponding to data within a user profile.