# US 10,764,431 B1 · Method for conversion and classification of data based on context

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

- Number: US 10,764,431 B1
- Document ID: US10764431B1
- Status: Granted
- Kind: B1
- Application number: 16/573,110
- Filed: 2019-09-17
- Granted: 2020-09-01
- Earliest claimed: 2019-09-17
- Assignees: Capital One Services, LLC
- Applicants: Capital One Services, LLC
- Inventors: Vincent Pham; Fardin Abdi Taghi Abad; Reza Farivar; Kate Key; Anh Truong; Jeremy Goodsitt; Austin Walters
- Web page: https://patentgrep.com/patent/US10764431B1

## CPC

- H04M 3/42221: {Conversation recording systems (at the subscriber's set H04M1/656)}
- G06F 16/3329: Natural language query formulation
- G06F 16/355: Creation or modification of classes or clusters
- G06F 16/685: {using automatically derived transcript of audio data, e.g. lyrics (speech recognition G10L15/00)}
- G06N 20/00: Machine learning
- G10L 15/26: Speech to text systems (G10L15/08 takes precedence)
- H04M 3/42042: {Notifying the called party of information on the calling party (details within substation equipment H04M1/57, signalling details H04Q3/72)}
- H04M 3/5166: {in combination with interactive voice response systems or voice portals, e.g. as front-ends}
- H04M 3/5175: {Call or contact centers supervision arrangements}
- G06N 3/006: based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
- H04M 2201/18: Comparators
- H04M 2201/38: Displays
- H04M 2201/40: using speech recognition
- H04M 2203/2061: Language aspects
- H04M 2203/357: Autocues for dialog assistance
- H04M 2203/401: Performance feedback
- H04M 2203/551: Call history
- H04M: TELEPHONIC COMMUNICATION (circuits for controlling other apparatus via a telephone cable and not involving telephone switching apparatus G08)
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G10L: SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING

## Abstract

When a new communication is received at a server, the system can generate machine readable data pertaining to the communication, e.g., a transcript for the communication. Using the transcript, the system can utilize natural language processing techniques to find at least one request in the communication. The system can search summaries of prior communications, and if there is a communication including a similar request, the system can display a record of the prior communication.

## Claims

### Claim 1

A method comprising:
receiving a first phone call;
transcribing the first phone call to create a first transcript;
performing a natural language processing technique to determine a first discussion attribute of the first transcript;
associating the first transcript with a first discussion attribute label selected from a group of labels including a plurality of discussion attribute labels;
receiving a second phone call;
transcribing the second phone call to create a second transcript;
performing the natural language processing technique to determine a second discussion attribute of the second call;
associating the second transcript with a second discussion attribute label selected from the group of labels;
comparing the first discussion attribute label to the second discussion attribute label;
in response to a determination that the first discussion attribute label matches the second discussion attribute label, displaying a report of the first transcript to a recipient of the second call, wherein the report of the first transcript includes a summary of the first transcript;
searching a database including a plurality of transcripts, each transcript being associated with at least one discussion attribute label, wherein during the search, the second discussion attribute label is compared to the discussion attribute label of each of a subset of the plurality of transcripts;
selecting at least one of the plurality of transcripts which is associated with the discussion attribute label that matches the second discussion attribute label;
combining the report of the first transcript with a report of a selected transcript to create a combined report; and
displaying the combined report to the user.

### Claim 2 (depends on 1)

The method of claim 1, further comprising:
performing the natural language processing technique to determine a third discussion attribute of the second call;
associating the second transcript with a third discussion attribute label selected from the group of labels;
searching the database, wherein during the search, the third discussion attribute label is compared to the discussion attribute label of the plurality of transcripts;
in response to a determination that the discussion attribute label of one of the plurality of transcripts matches the third discussion attribute label, displaying a report of the one of the plurality of transcripts to the recipient of the second call.

### Claim 3 (depends on 1)

The method of claim 1, further comprising:
performing the natural language processing technique to determine another discussion attribute of the second call;
if the other discussion attribute of the second call matches the second discussion attribute of the second call, continue displaying the report of the first transcript to the recipient of the second call;
if the other discussion attribute of the second call does not match the second discussion attribute of the second call:
associating the second transcript with the other discussion attribute label selected from the group of labels;
searching the database, wherein during the search, the other discussion attribute label is compared to the discussion attribute label of the plurality of transcripts;
in response to a determination that the discussion attribute label of one of the plurality of transcripts matches the other discussion attribute label, display a report of the one of the plurality of transcripts to the recipient of the second call instead of the report of the first transcript.

### Claim 4 (depends on 1)

The method of claim 1, wherein the combined report includes a problem statement from the first transcript and a problem statement from the selected transcript.

### Claim 5 (depends on 4)

The method of claim 4, further comprising:
displaying a feedback prompt to the user regarding a quality of the combined report.

### Claim 6 (depends on 5)

The method of claim 5, further comprising:
storing the feedback in a database including a plurality of feedbacks relating to combined reports.

### Claim 7 (depends on 6)

The method of claim 6, further comprising:
training the machine learning model using the plurality of feedbacks as training data.

### Claim 8 (depends on 1)

The method of claim 1, wherein the combined report includes the summary from the first transcript and a summary from the selected transcript.

### Claim 9 (depends on 1)

The method of claim 1, wherein the combined report includes a duration from the first transcript and a duration from the selected transcript.

### Claim 10 (depends on 1)

The method of claim 1, wherein the combined report includes an identity of a caller from the first transcript and an identity of a caller from the selected transcript.

### Claim 11 (depends on 1)

The method of claim 1, wherein the natural language processing technique uses a machine learning model.

### Claim 12 (depends on 1)

The method of claim 1, wherein the first discussion attribute is a statement of a problem.

### Claim 13 (depends on 1)

The method of claim 1, wherein the report of the first transcript includes a duration of the call.

### Claim 14 (depends on 1)

The method of claim 1, wherein the report of the first transcript includes an identity of a participant in the call.

### Claim 15 (depends on 1)

The method of claim 1, wherein the report of the first transcript includes a statement of the problem.

### Claim 16 (depends on 1)

The method of claim 1, wherein a plurality of discussion attributes are associated with the first discussion attribute label.