Patent Issued for Fraudulent claims detection system (USPTO 11741478): United Services Automobile Association

2023 SEP 15 (NewsRx) -- By a News Reporter-Staff News Editor at Insurance Daily News -- From Alexandria, Virginia, NewsRx journalists report that a patent by the inventors Dixon, David Patrick (Boerne, TX, US), filed on May 28, 2020, was published online on August 29, 2023.

The patent’s assignee for patent number 11741478 is United Services Automobile Association (San Antonio, Texas, United States).

News editors obtained the following quote from the background information supplied by the inventors: “Insurance fraud is expensive for the insurance company and is often difficult to detect. People may voluntarily or involuntarily underinsure their property. Some people may attempt to get coverage after an incident occurs and then file a claim. Others may seek out ways to file a claim for a cash settlement while not actually sustaining a loss. Accordingly, it is desirable to improve the ability to detect fraudulent claims.

“The techniques introduced here may be better understood by referring to the following Detailed Description in conjunction with the accompanying drawings, in which like reference numerals indicate identical or functionally similar elements. Moreover, while the technology is amenable to various modifications and alternative forms, specific embodiments have been shown by way of example in the drawings and are described in detail below. The intention, however, is not to limit the technology to the particular embodiments described. On the contrary, the technology is intended to cover all modifications, equivalents, and alternatives falling within the scope of the technology as defined by the appended claims”

As a supplement to the background information on this patent, NewsRx correspondents also obtained the inventors’ summary information for this patent: “A company that sells insurance receives many calls and inquiries each year. The insurance company may use a call center system that can accept and facilitate a voice call and/or electronic chat with a user. In some cases, a user may ask questions to determine what insurance coverage they currently have and/or determine what coverage they are eligible for through a new policy. Insurance fraud is expensive and is currently difficult to identify. Insurance fraud may be attempted by new applicants or by users who already have a policy. Therefore, it is advantageous to have a system that can automatically identify possible fraud to avoid making unnecessary and unwarranted payments.

“To overcome at least some of the issues described above, this patent document describes systems, apparatus, and methods for automatically determining that some inquires and/or filed claims may be fraudulent. Questions and inquiries from a user can be analyzed by fraud analysis algorithms to identify a probability that a request may be fraudulent. The system can store analysis of the interactions and weighted scores that indicate the probability of fraud, and the information can be retrieved and matched to a subsequently filed insurance claim. Voice calls and chats can be stored and later matched with a filed insurance claim. The system can suggest questions for a customer service representative to ask during a call/chat with a user that can help identify fraud. After a claim is filed, the system can apply previously stored data and fraud analysis algorithms to determine a probability or likelihood that the claim is fraudulent and suggest a lower payout if the fraud is hard to prove. The example headings for the various sections below are used to facilitate the understanding of the disclosed subject matter and do not limit the scope of the claimed subject matter in any way. Accordingly, one or more features of one section can be combined with one or more features of another section.

“I. Automatically Detect Potential Fraud Before Claim Is Filed”

The claims supplied by the inventors are:

“1. A method for transferring communications to a fraud server, comprising: receiving, at a customer service representative device of an enterprise, a call or a chat associated with a claim from a user device associated with a user; collecting data associated with the call or the chat, wherein the data includes: content data from the call or the chat, a geolocation of the user device, and historical weather data associated with the geolocation of the user device; in response to receiving the call or the chat, initiating, by a fraud analysis machine-learning platform trained using at least one of stored potential fraud records and unmatched potential fraud records, a fraudulent claim detection process including: analyzing the data associated with the call or the chat to determine whether to identify the claim at the geolocation of the user device as fraudulent; determining a weighted level of fraud associated with the claim based on the historical weather data associated with the geolocation of the user device; comparing the weighted level of fraud to at least one weight threshold; identifying the user as a fraud risk based on the weighted level of fraud meeting or exceeding the at least one weight threshold; in response to the weighted level of fraud meeting or exceeding the at least one weight threshold, A) causing an indication of the fraud risk to be displayed on a display of the customer service representative device associated with the customer service representative, and B) transferring the call or the chat to a fraud server, and in response to transferring the call or the chat, sending, from the fraud server, a record of the call or the chat to a database for storage and retrieval and training the fraud server based on a subject of the call or the chat and a number of times a potential fraud record has been recorded in a stored user profile.

“2. The method of claim 1, further comprising generating a potential fraud alert based on information from the user, wherein the potential fraud alert is generated while the call or chat is active or after the call or chat has ended.

“3. The method of claim 1, further comprising: in response to the user being identified as an increased fraud risk, determining at least one question for a customer service representative of the enterprise to ask the user when the call or chat is active; and displaying the at least one question on the display associated with the customer service representative.

“4. The method of claim 1, wherein the data further includes an address of the user and historical weather data associated with the address, wherein the method further comprising: analyzing the data associated with the call or the chat to determine a second claim at the address of the user; and determining a second weighted level of fraud associated with the second claim based on the historical weather data associated with the address.

“5. The method of claim 1, wherein the determining further comprises increasing the weighted level of fraud when the data includes at least one word or phrase input by the user that is indicative of a hypothetical situation.

“6. The method of claim 1, wherein the determining further comprises increasing the weighted level of fraud when the data includes at least one word or phrase input by the user that is indicative of a request for a cash payment resulting from a potential claim.

“7. The method of claim 1, wherein the determining further comprises increasing the weighted level of fraud when the data includes at least one word or phrase input by the user that is indicative of time restrictions on filing a claim with respect to a policy issuance date.

“8. The method of claim 1, further comprising: identifying at least one chat session or voice recording associated with the user that includes questions related to the data associated with the call or chat, wherein the chat session or voice recording are stored in the database or memory before the call or chat; and in response to identifying the at least one chat session or voice recording, increasing the weighted level of fraud.

“9. The method of claim 1, further comprising: determining a user profile associated with the call or chat; determining a reason for the call or chat; determining activity associated with the user profile, the activity including at least one of a history of calls, history of chats or filed claims; and increasing the weighted level of fraud in response to determining that the activity is associated with the reason for the call or chat.

“10. A non-transitory machine-readable medium storing instructions that, when executed by a computing system, cause the computing fraud system to perform a process for transferring communications to a server, for the process comprising: receiving, at a customer service representative device of an enterprise, a call or a chat associated with a claim from a user device associated with the user; collecting data associated with the call or the chat, wherein the data includes: content data from the call or the chat, a geolocation of the user device, and historical weather data associated with the geolocation of the user device; in response to receiving the call or the chat, initiating, by a fraud analysis machine-learning platform trained using at least one of stored potential fraud records and unmatched potential fraud records, a fraudulent claim detection process including: analyzing the data associated with the call or the chat to determine whether to identify the claim at the geolocation of the user device as fraudulent; determining a weighted level of fraud associated with the claim based on the historical weather data associated with the geolocation of the user device; comparing the weighted level of fraud to at least one weight threshold; identifying the user as a fraud risk based on the weighted level of fraud meeting or exceeding the at least one weight threshold; in response to the weighted level of fraud meeting or exceeding the at least one weight threshold, A) causing an indication of the fraud risk to be displayed on a display of the customer service representative device associated with the customer service representative, and B) transferring the call or the chat to a fraud server; and in response to transferring the call or the chat, sending, from the fraud server, a record of the call or the chat to a database for storage and retrieval, and training the fraud server based on a subject of the call or the chat and a number of times a potential fraud record has been recorded in a stored user profile.

“11. The non-transitory machine-readable medium of claim 10, wherein the process further comprises: in response to the user being identified as an increased fraud risk, determining at least one question for a customer service representative of the enterprise to ask the user when the call or chat is active; and displaying the at least one question on a display of the customer service representative device associated with the customer service representative.

“12. The non-transitory machine-readable medium of claim 10, wherein the determining further comprises increasing the weighted level of fraud when the data includes at least one word or phrase input by the user that is indicative of a hypothetical situation.

“13. The non-transitory machine-readable medium of claim 10, wherein the analyzing further includes analyzing a tone of voice of the user.

“14. The non-transitory machine-readable medium of claim 10, wherein the process further comprises: identifying at least one chat session or voice recording associated with the user that includes questions related to the data associated with the call or chat, wherein the chat session or voice recording are stored in the database or memory before the call or chat; and in response to identifying the at least one chat session or voice recording, increasing the weighted level of fraud.

“15. The non-transitory machine-readable medium of claim 10, wherein the data further includes an address of the user and historical weather data associated with the address, wherein the process further comprises: analyzing the data associated with the call or the chat to determine a second claim at the address of the user; and determining a second weighted level of fraud associated with the second claim based on the historical weather data associated with the address.”

There are additional claims. Please visit full patent to read further.

For additional information on this patent, see: Dixon, David Patrick. Fraudulent claims detection system. U.S. Patent Number 11741478, filed May 28, 2020, and published online on August 29, 2023. Patent URL (for desktop use only): https://ppubs.uspto.gov/pubwebapp/external.html?q=(11741478)&db=USPAT&type=ids

(Our reports deliver fact-based news of research and discoveries from around the world.)