Method and apparatus for an AI-assisted virtual assistant

1 . A method for generating a personalized content suggestion in response to a query from a user, wherein the method is performed by a system comprising one or more processors, the method comprising:

receiving the query from the user;

generating one or more user parameters associated with the user, wherein the one or more user parameters comprise identifying information associated with the user;

determining a similarity between the query and data in a content repository, comprising obtaining a vector representation of the query, identifying one or more topic areas within the content repository based on the one or more user parameters, and comparing the vector representation of the query to a vector encoding the data in the data repository within the one or more topic areas;

classifying the query as unmatched in accordance with the similarity being below a minimum threshold;

in accordance with classifying the query as unmatched, revising the query to generate a revised query, wherein revising the query comprises:

automatically traversing a first portion of a decision tree based at least in part on the one or more user parameters, wherein the one or more user parameters pre-answer one or more nodes of the first portion of the decision tree to narrow the query without presenting questions from the decision tree to the user; and

traversing a second portion of the decision tree to further narrow the query, comprising presenting one or more questions from the second portion of the decision tree to the user, receiving input from the user responsive to the one or more questions, and narrowing the query based on the received input from the user responsive to questions from the decision tree;

determining a predicted content topic of the revised query; and

generating the content suggestion based on the predicted content topic and the one or more user parameters.

2 . The method of claim 1 , wherein revising the query comprises:

providing a scoping prompt to the user;

receiving a scoping response from the user in response to providing the scoping prompt to the user; and

revising the query based on the scoping response.

3 . The method of claim 1 , wherein revising the query comprises:

providing one or more topic suggestions to the user;

receiving a selected content topic of one or more potential content topics in response to providing the one or more topic suggestions to the user; and

revising the query based on the selected content topic.

4 . The method of claim 3 , wherein the one or more topic suggestions are generated based on information from a knowledge substrate.

5 . The method of claim 1 , wherein generating the one or more user parameters comprises:

providing one or more user context prompts to the user;

receiving one or more user context responses from the user in response to providing the one or more user context prompts to the user; and

generating the one or more user parameters based on the one or more user context responses.

6 . The method of claim 1 , wherein generating the one or more user parameters comprises analyzing metadata associated with the user.

7 . The method of claim 6 , wherein the metadata associated with the user comprises one or more of the user's search history, metadata associated with one or more team members on a team with the user, past projects completed by the user, and a job title of the user.

8 . The method of claim 1 , wherein determining the predicted content topic comprises:

providing a clarifying prompt to the user;

receiving a clarifying response from the user in response to providing the clarifying prompt to the user; and

determining the predicted content topic based on the clarifying response.

9 . The method of claim 1 , wherein determining the predicted content topic of the revised query is based in part on the generated one or more user parameters.

10 . The method of claim 1 , wherein determining the predicted content topic comprises generating a confidence score corresponding to a level of confidence that the predicted content topic is associated with the revised query.

11 . The method of claim 10 , comprising directing the user to a live agent if the confidence score is below a predetermined threshold.

12 . The method of claim 1 , comprising providing the content suggestion to the user.

13 . The method of claim 12 , wherein providing the content suggestion to the user comprises providing a link to a content repository comprising data corresponding to the content suggestion.

14 . The method of claim 12 , wherein providing the content suggestion to the user comprises providing information corresponding to the content suggestion from a content repository directly to the user.

15 . The method of claim 1 , wherein receiving the query comprises receiving information from the user via one of a voice input, a chat message in a chatbot, and an email.

16 . A system for generating a personalized content suggestion in response to a query from a user, the system comprising one or more processors configured to cause the system to:

receive the query from the user;

generate one or more user parameters associated with the user, wherein the one or more user parameters comprise identifying information associated with the user;

determine a similarity between the query and data in a content repository, comprising obtaining a vector representation of the query, identifying one or more topic areas within the content repository based on the one or more user parameters, and comparing the vector representation of the query to a vector encoding the data in the data repository within the one or more topic areas;

classify the query as unmatched in accordance with the similarity being below a minimum threshold;

in accordance with classifying the query as unmatched, revise the query to generate a revised query, wherein revising the query comprises:

automatically traversing a first portion of a decision tree based at least in part on the one or more user parameters, wherein the one or more user parameters pre-answer one or more nodes of the first portion of the decision tree to narrow the query without presenting questions from the decision tree to the user; and

traversing a second portion of the decision tree to further narrow the query, comprising presenting one or more questions from the second portion of the decision tree to the user, receiving input from the user responsive to the one or more questions, and narrowing the query based on the received input from the user responsive to questions from the decision tree;

determine a predicted content topic of the revised query; and

generate the content suggestion based on the predicted content topic and the one or more user parameters.

17 . A non-transitory computer-readable storage medium storing instructions for generating a personalized content suggestion in response to a query from a user, the instructions configured to be executed by a system comprising one or more processors to cause the system to:

receive the query from the user;

generate one or more user parameters associated with the user, wherein the one or more user parameters comprise identifying information associated with the user;

determine a similarity between the query and data in a content repository, comprising obtaining a vector representation of the query, identifying one or more topic areas within the content repository based on the one or more user parameters, and comparing the vector representation of the query to a vector encoding the data in the data repository within the one or more topic areas;

classify the query as unmatched in accordance with the similarity being below a minimum threshold;

in accordance with classifying the query as unmatched, revise the query to generate a revised query, wherein revising the query comprises:

automatically traversing a first portion of a decision tree based at least in part on the one or more user parameters, wherein the one or more user parameters pre-answer one or more nodes of the first portion of the decision tree to narrow the query without presenting questions from the decision tree to the user; and

traversing a second portion of the decision tree to further narrow the query, comprising presenting one or more questions from the second portion of the decision tree to the user, receiving input from the user responsive to the one or more questions, and narrowing the query based on the received input from the user responsive to questions from the decision tree;

determine a predicted content topic of the revised query; and

generate the content suggestion based on the predicted content topic and the one or more user parameters.