# US 2022/0191154 A1 · Vitrual-assistant-based resolution of user inquiries via failure-triggered document presentation

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2022/0191154 A1
- Document ID: US20220191154A1
- Status: Published application
- Kind: A1
- Application number: 17/118,478
- Filed: 2020-12-10
- Published: 2022-06-16
- Earliest claimed: 2020-12-10
- Assignees: Capital One Services, LLC
- Applicants: Capital One Services, LLC
- Inventors: Anh TRUONG; Austin WALTERS
- Web page: https://patentgrep.com/patent/US20220191154A1

## CPC

- H04L 51/02: using automatic reactions or user delegation, e.g. automatic replies or chatbot-generated messages
- G06F 3/167: {Audio in a user interface, e.g. using voice commands for navigating, audio feedback}
- G06F 40/30: Semantic analysis
- G06N 3/02: Neural networks
- H04L 51/046: Interoperability with other network applications or services
- H04L 51/18: Commands or executable codes
- G06N 3/084: Backpropagation, e.g. using gradient descent
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

In certain embodiments, document-based resolution of user inquiries may be facilitated. A predicted intent of the user is determined based on chat activity information associated with the user. A response to a user inquiry may be provided to the user via a chat interface based on the predicted intent. User response may be obtained for the response, which may indicate a failure in providing a resolution to the user (e.g., regarding a user inquiry). Upon detecting a failure in providing the resolution, a document associated with the user and matching the predicted intent may be obtained and presented to via the chat interface. The document may have content related to the predicted intent. The document may be presented as a response to the user response to seek a confirmation from the user regarding its relevance to the user inquiry.

## Claims

### Claim 1

A system for facilitating a virtual-assistant-based resolution of user inquiries via sentiment-triggered document presentation, the system comprising:
a computer system that comprises one or more processors programmed with computer program instructions that, when executed, cause operations comprising:
performing sentiment analysis monitoring on user inputs and responses obtained via a chat interface during a chat session between a virtual assistant and a user;
providing a first user input of the user obtained via the chat interface as input to a neural network to obtain a predicted intent of the user;
generating, via the chat interface, a first response related to the first user input based on the predicted intent;
in response to detecting, via the sentiment analysis monitoring, a negative sentiment in connection with a second user input of the user following the first response, selecting a document portion of a document associated with the user over one or more other document portions of the document based on the document portion being a greater match with the predicted intent than the one or more other document portions; and
generating, via the chat interface, an image depicting the selected document portion in connection with providing a second response related to the second user input, wherein the second response indicates a request to provide a confirmation related to the selected document portion.

### Claim 2 (depends on 1)

The system of claim 1, wherein the operations further comprise:
in response to the detected negative sentiment, providing a graphical user interface element in the chat interface for facilitating the user to upload one or more documents.

### Claim 3 (depends on 1)

The system of claim 1, wherein the operations further comprise:
obtaining metadata of each of a plurality of documents associated with the user based on content extracted from each of the documents; and
obtaining the document from the plurality of documents based on metadata of the document matching the predicted intent.

### Claim 4 (depends on 1)

The system of claim 1, wherein the operations further comprise:
providing the document to a second neural network to obtain concepts associated with document portions of the document, the concepts comprising one or more first concepts associated with the document portion and one or more second concepts associated with another document portion of the document; and
determining a first probability of the document portion matching the predicted intent based on the one or more first concepts and a second probability of the other document portion matching the predicted intent based on the one or more second concepts,
wherein selecting the document portion comprises selecting the document portion over the other document portion based on the first probability being greater than the second probability.

### Claim 5

A method comprising:
obtaining a first user input of a user via a chat interface during a chat session with the user;
determining a predicted intent of the user during the chat session based on the first user input;
generating, via the chat interface, a first response related to the first user input based on the predicted intent; and
in response to detecting a negative sentiment in connection with one or more second user inputs following the first response, obtaining a document associated with the user based on the predicted intent and providing one or more document portions of the document for presentation on the chat interface in connection with providing a second response related to the one or more second user inputs,
wherein providing the one or more document portions of the document includes:
providing the document to a prediction model to obtain concepts associated with document portions of the document, the concepts comprising one or more first concepts associated with the one or more document portions and one or more second concepts associated with one or more other document portions of the document;
determining one or more first probabilities of the one or more document portions matching the predicted intent based on the one or more first concepts and one or more second probabilities of the one or more other document portions matching the predicted intent based on the one or more second concepts; and
selecting the one or more document portions over the one or more other document portions based on the one or more first probabilities being greater than the one or more second probabilities.

### Claim 6

(canceled)

### Claim 7 (depends on 5)

The method of claim 5, wherein providing the one or more document portions includes:
presenting, based on the selection, an image depicting the one or more document portions.

### Claim 8 (depends on 5)

The method of claim 5, wherein providing the one or more document portions includes:
presenting the document via the chat interface; and
emphasizing, based on the selection, the one or more document portions over the one or more other document portions of the document during the presentation of the document.

### Claim 9

(canceled)

### Claim 10

(canceled)

### Claim 11 (depends on 5)

The method of claim 5, further comprising:
performing sentiment analysis monitoring on the one or more second user inputs to detect the negative sentiment.

### Claim 12 (depends on 5)

The method of claim 5, further comprising:
in response to detecting the negative sentiment, presenting a graphical user interface element in the chat interface for facilitating the user to upload a specified document.

### Claim 13 (depends on 5)

The method of claim 5, wherein obtaining the document includes:
obtaining metadata of each of a plurality of documents associated with the user based on content extracted from each of the documents, and
obtaining the document from the plurality of documents based on metadata of the document matching the predicted intent.

### Claim 14 (depends on 5)

The method of claim 5, wherein determining the predicted intent includes:
obtaining an intent set using a prediction model, wherein each predicted intent of the intent set is associated with a probability of matching a current intent of the user, and
selecting the predicted intent from the intent set based on the predicted intent being associated with a higher probability than other predicted intents of the intent set.

### Claim 15

(canceled)

### Claim 16 (depends on 5)

The method of claim 5, wherein obtaining the document includes:
generating the document based on a plurality of documents associated with the user.

### Claim 17

A non-transitory computer-readable medium comprising instructions that, when executed by one or more processors, cause operations comprising:
obtaining a first user input of a user via a chat interface during a chat session with the user;
determining a predicted intent of the user during the chat session based on the first user input;
generating, via the chat interface, a first response related to the first user input based on the predicted intent; and
in response to detecting a negative sentiment in connection with one or more second user inputs following the first response, obtaining a document associated with the user based on the predicted intent and providing the document for presentation on the chat interface in connection with providing a second response related to the one or more second user inputs,
wherein providing the document includes:
providing the document to a prediction model to obtain concepts associated with document portions of the document, the concepts comprising one or more first concepts associated with one or more document portions and one or more second concepts associated with one or more other document portions of the document;
determining one or more first probabilities of the one or more document portions matching the predicted intent based on the one or more first concepts and one or more second probabilities of the one or more other document portions matching the predicted intent based on the one or more second concepts; and
selecting the one or more document portions over the one or more other document portions based on the one or more first probabilities being greater than the one or more second probabilities.

### Claim 18 (depends on 17)

The computer-readable medium of claim 17, wherein providing the document includes:
presenting the document via the chat interface; and
emphasizing, based on the selection, the one or more document portions over the one or more other document portions of the document during the presentation of the document.

### Claim 19 (depends on 17)

The computer-readable medium of claim 17, wherein providing the document includes:
presenting, based on the selection, an image depicting the one or more document portions.

### Claim 20 (depends on 17)

The computer-readable medium of claim 17, the operations further comprising:
performing sentiment analysis monitoring on the one or more second user inputs to detect the negative sentiment.

### Claim 21 (depends on 17)

The computer-readable medium of claim 17, wherein obtaining the document includes:
obtaining metadata of each of a plurality of documents associated with the user based on content extracted from each of the documents, and
obtaining the document from the plurality of documents based on metadata of the document matching the predicted intent.

### Claim 22

(canceled)

### Claim 23

A method comprising:
obtaining a first user input of a user via a chat interface during a chat session with the user;
determining a predicted intent of the user during the chat session based on the first user input;
generating, via the chat interface, a first response related to the first user input based on the predicted intent; and
in response to detecting a negative sentiment in connection with one or more second user inputs following the first response, obtaining a document associated with the user based on the predicted intent and providing one or more document portions of the document for presentation on the chat interface in connection with providing a second response related to the one or more second user inputs,
wherein providing the one or more document portions of the document includes:
presenting the document via the chat interface; and
emphasizing, based on the selection, the one or more document portions over the one or more other document portions of the document during the presentation of the document.

### Claim 24 (depends on 23)

The method of claim 23, wherein providing the one or more document portions includes:
presenting, based on the selection, an image depicting the one or more document portions.

### Claim 25

A non-transitory computer-readable medium comprising instructions that, when executed by one or more processors, cause operations comprising:
obtaining a first user input of a user via a chat interface during a chat session with the user;
determining a predicted intent of the user during the chat session based on the first user input;
generating, via the chat interface, a first response related to the first user input based on the predicted intent; and
in response to detecting a negative sentiment in connection with one or more second user inputs following the first response, obtaining a document associated with the user based on the predicted intent and providing the document for presentation on the chat interface in connection with providing a second response related to the one or more second user inputs,
wherein providing the document includes:
selecting, for presentation via the chat interface, one or more document portions of the document over one or more other document portions based on the one or more document portions being a greater match with the predicted intent than the one or more other document portions;
presenting the document via the chat interface; and
emphasizing, based on the selection, the one or more document portions over the one or more other document portions of the document during presentation of the document.

### Claim 26 (depends on 25)

The computer-readable medium of claim 25, wherein providing the document includes:
presenting, based on the selection, an image depicting the one or more document portions.
