Researchers Submit Patent Application, “Machine Learning Technologies for Efficiently Obtaining Insurance Coverage”, for Approval (USPTO 20220215476): Patent Application

2022 JUL 21 (NewsRx) -- By a News Reporter-Staff News Editor at Insurance Daily News -- From Washington, D.C., NewsRx journalists report that a patent application by the inventors Frankowiak, Sara (Bloomington, IL, US); Isaacs, Craig Dean (Bloomington, IL, US); McCarty, Jeff (Bloomington, IL, US); Roll, Leif Agerholm (Bloomington, IL, US); Taylor, Kelli (Bloomington, IL, US), filed on June 24, 2020, was made available online on July 7, 2022.

No assignee for this patent application has been made.

News editors obtained the following quote from the background information supplied by the inventors: “Individuals who seek insurance coverage and are sensitive to pricing and product features (e.g., coverage types and/or limits, deductibles, etc.), or “frequent shoppers,” often expend considerable time and effort in finding insurance providers that best meet their needs. Conventionally, a frequent shopper finds an insurance provider by way of an agent/broker, an aggregator, a comparison web site, general web browsing, etc. Once the frequent shopper obtains an insurance policy from the desired provider, the frequent shopper is typically tied to that provider, and to the rate and product features of the policy offered by the provider, until and unless he or she proactively shops around for a new provider offering a policy with a better rate and/or product features. For example, a frequent shopper might decide to look into the offerings of other insurance providers when the frequent shopper’s current policy is up for renewal. Thus, a frequent shopper typically must either spend time and effort looking for a better-priced insurance offering on a recurring basis (e.g., once every six months or annually), or simply renew his or her current policy regardless of whether that policy provides the best rate and/or product features. Conventional agency-based insurance models may not suffice to meet a frequent shopper’s needs, due to the perceived additional cost associated with having an agent.”

As a supplement to the background information on this patent application, NewsRx correspondents also obtained the inventors’ summary information for this patent application: “The present embodiments may, inter alia, automatically provide frequent shoppers with insurance policies that offer superior rates and/or product features on a continuing basis (e.g., across multiple policy terms), thereby reducing or eliminating the time and/or effort that frequent shoppers must spend researching the offerings of different insurance providers, as well as providing frequent shoppers with insurance policies that have lower cost and/or are more reflective of a risk score, characteristics, and/or preferences of the frequent shopper as they change over time. The terms “frequent shopper,” “consumer,” and “customer” are utilized interchangeably herein, and generally refer to a person who is an insured party or a potential insured party, regardless of how frequently that individual in fact would like to shop for insurance coverage or has shopped for insurance coverage in the past. A frequent shopper may be represented by himself or herself, or may be represented by an agent (e.g., by a spouse, a person who has power of attorney for the frequent shopper, an administrative assistant, etc.).

“An intermediary entity may act on behalf of frequent shoppers and/or their agents (i.e., with the consent of the frequent shoppers and/or agents) to find policy rates and/or other features that best meet the frequent shoppers’ insurance requirements and/or preferences. Based upon an analysis of individual frequent shopper characteristics and/or insurance preferences, each individual frequent shopper may be grouped with other insurance frequent shoppers that have the same or similar characteristics and/or insurance preferences. These “affinity groupings” (or “affinity groups”) may be based upon demographic information for the frequent shoppers (e.g., gender, birth date, etc.), information about property of the frequent shoppers (e.g., a make, model and year of an automobile, etc.), claim and/or accident history of the frequent shoppers, risk (or lack thereof) characteristics of the frequent shoppers, insurance claim expectations of the frequent shoppers, insurance company ratings, the content and/or availability of telematics data obtained from vehicles and/or mobile devices of the frequent shoppers, driving behaviors of the frequent shoppers, etc.

“The right to provide insurance coverage for the affinity groupings (either on a per-member basis or to each affinity group as a whole) may be offered for sale to various insurance providers, such as through an online auction. In some embodiments, other entities may also participate in the online auction. For example, reinsurers and/or entities that manage investment funds (e.g., hedge fund companies seeking arbitrage opportunities) may participate in the online auction, e.g., if those entities have agreements with insurance providers are licensed to write insurance and can legally service claims, etc., for the members of the affinity group.

“Once a winning bid is accepted, any existing insurance policies of the frequent shoppers affiliated with the auctioned group may (or may not) be updated to reflect new insurance policy terms or parameters (e.g., premiums, rates, etc.), discounts, refunds, etc. In some cases, new insurance policies may be provided to one or more frequent shoppers (such as when a frequent shopper is an insurance applicant, or when an existing insurance policy is canceled and a new policy is issued in its stead). The affinity groups may be updated (and/or new affinity groups may be created) over time as new or more recent frequent shopper characteristic data and/or preference information is collected and/or updated. The insurance policies associated with the updated (or new) affinity groups may then be re-auctioned (or auctioned).

“Additionally or alternatively, insurance providers may mitigate the risks associated with insurance policies that are already in effect (or will soon be in effect), by grouping/segmenting those policies and auctioning the opportunity to reinsure those policy groups to other entities (e.g., reinsurers). The grouping for these auctions may correspond to the affinity groups discussed above (e.g., with a particular group of policies that is being auctioned consisting of the policies of all members of a particular affinity group), or may be an independent/subsequent grouping of the insurance policies, for example.

“Various machine learning technologies described herein may increase the efficiency of any or all of the grouping and/or auctioning techniques discussed above, in some embodiments. For example, machine learning models may be used to evaluate risks (e.g., determine risk scores/classifications and/or infer risk-related characteristics) associated with different frequent shoppers for a particular type of insurance (e.g., risks of vehicular accidents and/or theft for auto insurance), prior to segmenting those frequent shoppers into different affinity groups based upon those risks. Machine learning may also be used to define and/or update/refine criteria for different affinity groups, e.g., by using regression models to determine which groupings of frequent shoppers have historically attracted more interest (e.g., more frequent and/or higher bids) from insurance providers, and/or have historically had more stable group membership, etc.

“Machine learning techniques may also, or instead, be used to set up an auction, and/or to facilitate the auction itself. For example, machine learning models may help determine which insurance providers to invite to participate in an auction, by predicting which providers are more likely to be interested in (e.g., more likely to submit bids for) providing insurance coverage to a particular affinity group, and/or may determine a suitable starting bid (e.g., “reserve”) amount, etc.

“Many or all facets of the auction process, and/or other procedures prior to and/or after the auction process, may be automated. For example, communications with auction participants (e.g., insurance providers and/or reinsurers, etc.), and/or communications with consumers that occur before, during, and/or after the auction process, may be automated. For example, notifying insurance providers and/or reinsurers regarding an upcoming auction, communicating bid amounts among providers and/or reinsurers during an auction, notifying auction winners, corresponding with consumers regarding insurance provider placements, billings, sending insurance cards, etc., and/or other communications may be automated. In some embodiments, records associated with consumers, insurance providers (and/or reinsurers or other auction participants), affinity groups, and/or auctions may be securely stored utilizing blockchain systems and/or techniques.

“In one aspect, a computer-implemented method comprises: (1) dividing, by one or more processors, a plurality of consumers into multiple affinity groups based upon one or more characteristics and one or more preferences of the plurality of consumers, at least in part by analyzing the one or more characteristics and/or the one or more preferences of the plurality of consumers using a machine learning model; (2) auctioning, by the one or more processors and via a communications network, an opportunity to provide insurance for one or more of the multiple affinity groups; (3) receiving, by the one or more processors and via the communications network, one or more bids for purchase and/or offers of insurance for the one or more of the multiple affinity groups; (4) accepting, by the one or more processors, a winning bid of the one or more bids; and/or (5) causing, by the one or more processors, individual insurance policies or a group insurance policy to be provided to or updated for consumers associated with a particular affinity group corresponding to the winning bid, thereby providing lower cost insurance and/or insurance that is more reflective of actual risk, or lack thereof, to the consumers associated with the particular affinity group.

“In another aspect, a system comprises a persistent memory storing a consumer profile database, a communication interface configured to communicate with remote devices via a communications network, one or more processors, and/or one or more non-transitory, computer-readable media storing instructions. The instructions, when executed by the one or more processors, cause the system to: (1) divide a plurality of consumers into multiple affinity groups based upon one or more characteristics and one or more preferences of the plurality of consumers, at least in part by analyzing the one or more characteristics and/or the one or more preferences of the plurality of consumers using a machine learning model; (2) auction, via the communication interface and the communications network, an opportunity to provide insurance for one or more of the multiple affinity groups; (3) receive, via the communication interface and the communications network, one or more bids for purchase and/or offers of insurance for the one or more of the multiple affinity groups; (4) accept a winning bid of the one or more bids; and/or (5) cause individual insurance policies or a group insurance policy to be provided to or updated for consumers associated with a particular affinity group corresponding to the winning bid, thereby providing lower cost insurance and/or insurance that is more reflective of actual risk, or lack thereof, to the consumers associated with the particular affinity group.

“Advantages will become more apparent to those skilled in the art from the following description of the preferred embodiments which have been shown and described by way of illustration. As will be realized, the present embodiments may be capable of other and different embodiments, and their details are capable of modification in various respects. Accordingly, the drawings and description are to be regarded as illustrative in nature and not as restrictive.”

There is additional summary information. Please visit full patent to read further.”

The claims supplied by the inventors are:

“1. A computer-implemented method comprising: dividing, by one or more processors, a plurality of consumers into multiple affinity groups based at least upon one or more characteristics and one or more preferences of the plurality of consumers, at least in part by analyzing the one or more characteristics and/or the one or more preferences of the plurality of consumers using a machine learning model; auctioning, by the one or more processors and via a communications network, an opportunity to provide insurance for one or more of the multiple affinity groups; receiving, by the one or more processors and via the communications network, one or more bids for purchase and/or offers of insurance for the one or more of the multiple affinity groups; accepting, by the one or more processors, a winning bid of the one or more bids; and causing, by the one or more processors, individual insurance policies or a group insurance policy to be provided to or updated for consumers associated with a particular affinity group corresponding to the winning bid, thereby providing lower cost insurance and/or insurance that is more reflective of actual risk, or lack thereof, for the consumers associated with the particular affinity group.

“2. The computer-implemented method of claim 1, wherein dividing the plurality of consumers into multiple affinity groups includes: determining risk scores for the plurality of consumers by analyzing the one or more characteristics of the plurality of consumers using the machine learning model; and dividing the plurality of consumers into the multiple affinity groups based at least upon the risk scores and the one or more preferences of the plurality of consumers.

“3. The computer-implemented method of claim 1, wherein dividing the plurality of consumers into multiple affinity groups includes: determining preference classifications for the plurality of consumers by analyzing the one or more preferences of the plurality of consumers using the machine learning model; and dividing the plurality of consumers into the multiple affinity groups based at least upon the preference classifications and the one or more characteristics of the plurality of consumers.

“4. The computer-implemented method of claim 1, wherein dividing the plurality of consumers into multiple affinity groups includes: determining classifications for the plurality of consumers by analyzing the one or more characteristics and the one or more preferences of the plurality of consumers using the machine learning model; and dividing the plurality of consumers into the multiple affinity groups based at least upon the classifications.

“5. The computer-implemented method of claim 1, wherein dividing the plurality of consumers into multiple affinity groups includes: using the machine learning model to infer at least one additional characteristic and/or at least one additional preference for at least some of the plurality of consumers; and dividing the plurality of consumers into the multiple affinity groups based at least in part upon the at least one additional characteristic and/or the at least one additional preference.

“6. The computer-implemented method of claim 1, further comprising, prior to dividing the plurality of consumers into the multiple affinity groups: determining, by the one or more processors analyzing historical data indicative of (i) consumer characteristics and/or preferences for different affinity groups and (ii) bidding activity for the different affinity groups, requirements for membership in each of the multiple affinity groups.

“7. The computer-implemented method of claim 1, wherein the machine learning model is a neural network, and further comprising, prior to dividing the plurality of consumers into the multiple affinity groups: training the neural network using historical data indicative of (i) consumer characteristics and/or preferences for different consumers and (ii) risk-related outcomes for the different consumers.

“8. The computer-implemented method of claim 1, wherein auctioning the opportunity to provide insurance for one or more of the multiple affinity groups includes: for each affinity group of the one or more of the multiple affinity groups, auctioning the opportunity to provide individual insurance policies for each consumer within the affinity group.

“9. The computer-implemented method of claim 1, wherein auctioning the opportunity to provide insurance for one or more of the multiple affinity groups includes: for each affinity group of the one or more of the multiple affinity groups, auctioning the opportunity to provide a group insurance policy for all consumers within the affinity group.

“10. The computer-implemented method of claim 1, further comprising, prior to dividing the plurality of consumers into the multiple affinity groups: receiving, by the one or more processors, vehicle telematics data for the plurality of consumers, the vehicle telematics data being indicative of one or more driving behaviors, and the one or more characteristics of the plurality of consumers including the one or more driving behaviors.

“11. A system comprising: a persistent memory storing a consumer profile database; a communication interface configured to communicate with remote devices via a communications network; one or more processors; and one or more non-transitory, computer-readable media storing instructions that, when executed by the one or more processors, cause the system to divide a plurality of consumers into multiple affinity groups based at least upon one or more characteristics and one or more preferences of the plurality of consumers that are included in the consumer profile database, at least in part by analyzing the one or more characteristics and/or the one or more preferences of the plurality of consumers using a machine learning model, auction, via the communication interface and the communications network, an opportunity to provide insurance for one or more of the multiple affinity groups, receive, via the communication interface and the communications network, one or more bids for purchase and/or offers of insurance for the one or more of the multiple affinity groups, accept a winning bid of the one or more bids, and cause individual insurance policies or a group insurance policy to be provided to or updated for consumers associated with a particular affinity group corresponding to the winning bid, thereby providing lower cost insurance and/or insurance that is more reflective of actual risk, or lack thereof, for the consumers associated with the particular affinity group.

“12. The system of claim 11, wherein dividing the plurality of consumers into multiple affinity groups includes: determining risk scores for the plurality of consumers by analyzing the one or more characteristics of the plurality of consumers using the machine learning model; and dividing the plurality of consumers into the multiple affinity groups based at least upon the risk scores and the one or more preferences of the plurality of consumers.

“13. The system of claim 11, wherein dividing the plurality of consumers into multiple affinity groups includes: determining preference classifications for the plurality of consumers by analyzing the one or more preferences of the plurality of consumers using the machine learning model; and dividing the plurality of consumers into the multiple affinity groups based at least upon the preference classifications and the one or more characteristics of the plurality of consumers.

“14. The system of claim 11, wherein dividing the plurality of consumers into multiple affinity groups includes: determining classifications for the plurality of consumers by analyzing the one or more characteristics and the one or more preferences of the plurality of consumers using the machine learning model; and dividing the plurality of consumers into the multiple affinity groups based at least upon the classifications.

“15. The system of claim 11, wherein dividing the plurality of consumers into multiple affinity groups includes: using the machine learning model to infer at least one additional characteristic and/or at least one additional preference for at least some of the plurality of consumers; and dividing the plurality of consumers into the multiple affinity groups based at least in part upon the at least one additional characteristic and/or the at least one additional preference.

“16. The system of claim 11, wherein the instructions further cause the system to, prior to dividing the plurality of consumers into the multiple affinity groups: determine, by analyzing historical data indicative of (i) consumer characteristics and/or preferences for different affinity groups and (ii) bidding activity for the different affinity groups, requirements for membership in each of the multiple affinity groups.

“17. The system of claim 11, wherein the machine learning model is a neural network, and wherein the instructions further cause the system to, prior to dividing the plurality of consumers into the multiple affinity groups: train the neural network using historical data indicative of (i) consumer characteristics and/or preferences for different consumers and (ii) risk-related outcomes for the different consumers.

“18. The system of claim 11, wherein auctioning the opportunity to provide insurance for one or more of the multiple affinity groups includes: for each affinity group of the one or more of the multiple affinity groups, auctioning the opportunity to provide individual insurance policies for each consumer within the affinity group.

“19. The system of claim 11, wherein auctioning the opportunity to provide insurance for one or more of the multiple affinity groups includes: for each affinity group of the one or more of the multiple affinity groups, auctioning the opportunity to provide a group insurance policy for all consumers within the affinity group.”

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

For additional information on this patent application, see: Frankowiak, Sara; Isaacs, Craig Dean; McCarty, Jeff; Roll, Leif Agerholm; Taylor, Kelli. Machine Learning Technologies for Efficiently Obtaining Insurance Coverage. Filed June 24, 2020 and posted July 7, 2022. Patent URL: https://appft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PG01&p=1&u=%2Fnetahtml%2FPTO%2Fsrchnum.html&r=1&f=G&l=50&s1=%2220220215476%22.PGNR.&OS=DN/20220215476&RS=DN/20220215476

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