Supplemental review process determination utilizing advanced analytics decision making model

• US 10,706,474 B2
• Filed: 04/25/2016
• Issued: 07/07/2020
• Est. Priority Date: 04/25/2016
• Status: Active Grant

• Current Assignee: Hartford Fire Insurance Company (The Hartford Financial Services Group, Inc.)
• Original Assignee: Hartford Fire Insurance Company (The Hartford Financial Services Group, Inc.)
• Inventors: Rieth, Kimberly A., Roberts, Lucas Raymond, Sheaffer, Christopher L.
• Primary Examiner(s): Kalinowski, Alexander G
• Assistant Examiner(s): Lickteig, Blane A
• Application Number:
  US15/137,667
• Publication Number:
  US 20170308958A1
• Time in Patent Office: 1,534 Days
• Field of Search: None
• CPC Class Codes: G06Q 40/08 Insurance

Abstract

A system may automatically identify electronic records to be routed to a supplemental review process via an automated back-end application computer server. The system may include a data store containing electronic records, each electronic record representing a potential new risk association with an entity, and each electronic record may contain a record identifier and a set of record characteristic values. The computer server may then access the electronic records and automatically generate, by an advanced analytics decision making model based on the record characteristic values, a supplemental review process indication for the potential new risk association of each entity. Indications representing the supplemental review process indications may be transmitted in connection with an interactive user interface display and it may be arranged for electronic records having supplemental review process indications that meet a pre-determined condition to be automatically routed such that those electronic records will undergo the supplemental review process.

Claims

1. A system to automatically identify electronic records to be routed to a supplemental insurance review process comprising:

(a) a data store containing a plurality of historical electronic records comprising determinate historical data originating from determinate data sources, and indeterminate historical data originating from indeterminate data sources, and a batch of electronic records representing a potential new risk association with an entity, wherein each record in the batch of electronic records contains a record identifier and a set of record characteristic values including at least;

one record characteristic value representing a characteristic of the entity,

one record characteristic value representing an original source of the each record, and

one record characteristic value representing third-party data;

(b) an application computer server coupled to the data store including at least one processor and a non-tangible, computer readable medium storing programmed instructions configured to;

extract the indeterminate historical data from raw files;

store the indeterminate historical data, the determinate historical data, and the data representing potential new risk associations in the data store;

train a plurality of independent analytics decision making models using the stored indeterminate and determinate historical data, wherein the plurality of trained independent analytics decision making models apply statistical techniques including at least machine learning, and wherein at least one of the plurality of independent analytics decision making models is associated with a probability, p, of an electronic record representing a potential new risk undergoing a supplemental review process defined as;

2. The system of claim 1, wherein at least one of the plurality of independent analytics decision making models is associated with at least one of:

(i) different weighing values for the same record characteristic, (ii) different record characteristic values, and (iii) different types of predictive models.

3. The system of claim 2, wherein the model aggregation and data fusion component assigns a first priority value to a first analytics decision making model and a second priority value to a second analytics decision making model, and said combining is based at least in part on the first and second priority values.

4. The system of claim 1, wherein the score value for the each record is not displayed via the interactive user interface display.

5. The system of claim 1, wherein the each record is associated with a potential insured the potential new risk association is a potential new insurance policy, and the supplemental review process comprises an underwriting evaluation of the potential new insurance policy.

6. The system of claim 5, wherein the at least one record characteristic value representing a characteristic of the entity includes at least three of:

(i) location information, (ii) an industry, (iii) a presence of a general industry classification, (iv) a customer score, (v) a number of locations, (vi) a number of states, (vii) a presence of one or more restricted terms, and (viii) a presence of one or more encouraged terms.

7. The system of claim 5, wherein the at least one record characteristic value representing an original source of the electronic record includes at least two of:

(i) an identifier of the original source from which the potential new insurance policy was received, (ii) an insurance agency score, (iii) a cancel rate, (iv) a new-business cancel rate, (v) a back-dating rate, (vi) a declination rate, (v a return on equity value, and (viii) an ultimate cost ratio.

8. The system of claim 5, wherein there the at least one record characteristic value representing third-party data includes at least two of:

(i) a business credit score, (ii) a comparison of a third-party industry classification code with an internally generated classification code, (iii) social media information, (iv) economic indicators, and (v) macro-economic conditions.

9. The system of claim 3, wherein at least some of the set of record characteristic values are associated with other types of insurance, and the supplemental review process indication is generated in accordance with a presence of at least one of:

(i) workers'"'"' compensation insurance, (ii) disability insurance, and (iii) general liability insurance.

10. The system of claim 1, wherein the each electronic record is associated with a record identifier and a communication address, and the sets of record characteristic values are collected via at least one of:

(i) sending a communication to a communication address and receiving, from a party associated with an electronic record having that communication address, a response to the communication, (ii) a postal mailing automatically generated by a distribution center, (iii) a postal mailing received by the distribution center, (iv) an email automatically generated by an email server, (v) information provided a web interface, (vi) an interactive voice response system associated with a telephone call center, (vii) a chat application that interacts with a party in real time, and (viii) a video link.

11. The system of claim 1, wherein the performance outcomes are associated with at least one of:

(i) manual underwriting decisions, (ii) a direct notice of cancellation, (iii) a do not renew decision, (iv) endorsement activity, and (v) a level of verifiability.

12. A method to automatically identify electronic records to be routed to a supplemental insurance review process, the method performed via at least one processor and a non-tangible, computer-readable medium storing programmed instructions in an application computer server, the method comprising:

extracting indeterminate historical data from raw files;

storing, in a data store, the indeterminate historical data, determinate historical data originating from determinate data sources, and a batch of electronic records representing a potential new risk association with an entity, wherein each record in the batch of electronic records contains a record identifier and a set of record characteristic values including at least;

one record characteristic value representing a characteristic of the entity,

one record characteristic value representing an original source of the each record, and

one record characteristic value representing third-party data;

training a plurality of independent analytics decision making models using the stored indeterminate and determinate historical data, wherein the plurality of trained independent analytics decision making models apply statistical techniques including at least machine learning, and wherein at least one of the plurality of independent analytics decision making models is associated with a probability, p, of an electronic record representing a potential new risk undergoing a supplemental review process defined as;

13. The method of claim 12, wherein each entity is associated with a potential insured and the potential new risk association is a potential new insurance policy and the supplemental review process comprises an underwriting evaluation of the potential new insurance policy.

14. A non-tangible, computer-readable medium of an automated back-end application computer server, the medium storing instructions that, when executed by the server, cause the processor to perform a method to automatically identify electronic records to be routed to a supplemental insurance review process, the method comprising:

extracting indeterminate historical data from raw files;

storing, in a data store, the indeterminate historical data, determinate historical data originating from determinate data sources, and a batch of electronic records representing a potential new risk association with an entity, wherein each record in the batch of electronic records contains a record identifier and a set of record characteristic values including at least;

one record characteristic value representing a characteristic of the entity,

one record characteristic value representing an original source of the each record, and

one record characteristic value representing third-party data;

training a plurality of independent analytics decision making models using the stored indeterminate and determinate historical data, wherein the plurality of trained independent analytics decision making models apply statistical techniques including at least machine learning, and wherein at least one of the plurality of independent analytics decision making models is associated with a probability, p, of an electronic record representing a potential new risk undergoing a supplemental review process defined as;

15. The medium of claim 14, wherein each entity is associated with a potential insured and the potential new risk association is a potential new insurance policy and the supplemental review process comprises an underwriting evaluation of the potential new insurance policy.