Insurance claims processing

• US 9,299,108 B2
• Filed: 02/22/2013
• Issued: 03/29/2016
• Est. Priority Date: 02/24/2012
• Status: Active Grant

• Current Assignee: TATA Consultancy Services Limited (Tata Sons Pvt Ltd.)
• Original Assignee: TATA Consultancy Services Limited (Tata Sons Pvt Ltd.)
• Inventors: Diana, Frank, Mulcahy, Kevin, Menkin, Gary, Bhargava, Ajay, Gupta, Ruchi, Bhattacharyya, Tumpa
• Primary Examiner(s): Poinvil, Frantzy
• Application Number:
  US13/774,027
• Publication Number:
  US 20130226623A1
• Time in Patent Office: 1,131 Days
• Field of Search: 705 3- 44
• US Class Current: 1/1
• CPC Class Codes: G06Q 10/10 Office automation; Time management
• CPC Class Codes: G06Q 40/08 Insurance

Abstract

Systems and methods for insurance claims processing in an insurance industry are described. The method comprises combining extracted claims data from one or more data sources to obtain a consolidated claims record and removing noise from text data of the consolidated claims record to obtain a claim dataset. The claims data comprises a plurality of claims. Further, ascertaining one or more suspicion indicators in the plurality of claims based on an analytical technique. Further, assigning a score to each of the plurality of claims based on at least one scoring rule. The score is an indicative of a level of suspicion of a claim. Furthermore, detecting at least one of insurance claims fraud and subrogation potential claims based on the score assigned to each of the plurality of claims.

Claims

1. A system for identifying insurance claims fraud and subrogation potential claims, the system comprising:

a processor;

a memory coupled to the processor, wherein the processor executes a plurality of modules stored in the memory, and wherein the plurality of modules comprising a data acquisition module configured to,

extract claims data from one or more data source, wherein the claims data includes structured data and unstructured data, and wherein the unstructured data includes imaged content and speech data;

convert the imaged content and the speech data to a textual format, wherein the imaged content includes photos and scanned copies of claim related information;

combine the claims data, having the textual format of the imaged content and the speech data, to obtain a consolidated claims record having text data, wherein data, from the claims data, in a language different from a language used in most of the claims data is translated to the language used in most of the claims data; and

remove noise from the text data of the consolidated claims records to obtain a claim dataset, wherein the claim dataset includes a plurality of claims;

an analysis module configured to ascertain one or more suspicion indicators in the plurality of claims based on concepts obtained from a fraud dictionary using one or more analytical technique, wherein the concepts are at least one of words, phrases, and entities that indicate suspicious claims; and

a scoring module configured to,

assign a score to each of the plurality of claims based on at least one scoring rule comprising one or more scoring parameters, wherein the score is indicative of a level of suspicion of a claim; and

identify at least one of insurance claims fraud and subrogation potential claims based on the score assigned to each of the plurality of claims.

2. The system as claimed in claim 1, wherein the noise comprises at least one of grammatical mistake, spelling mistakes, and punctuation mistakes.

3. The system as claimed in claim 1, wherein the one or more analytic technique is comprises a text analysis, a predictive model analysis, a link analysis, a business rules analysis, and a social-media analysis, and wherein the text analysis uses at least one text pattern to identify the suspicion indicators, and wherein the link analysis uses suspicious links between entities of two different claims to identify the suspicion indicators, and wherein the business rules analysis uses business rules to identify the suspicion indicators, and wherein the social-media analysis uses social-media content and social-media links between entities involved in a claim, from the among the plurality of claims, to identify the suspicion indicators in the claim.

4. The system as claimed in claim 1, wherein the scoring module is configured to create the at least one scoring rule for a specific client and a line of business (LOB).

5. The system as claimed in claim 1, wherein the fraud dictionary includes the at least one text pattern, the business rules, rules for identifying the suspicious links between entities of two different claims, and rules for identifying the social-media content and the social-media links.

6. The system as claimed in claim 1, wherein the scoring module is configured to update the dictionary based on a line of business (LOB).

7. The system as claimed in claim 1, wherein the scoring module is configured to recalculate the score upon receiving new information related to the plurality of claims.

8. The system as claimed in claim 1, wherein the scoring module is configured to categorize the plurality of claims into one or more categories based on the score assigned to each of the plurality of claims.

9. The system as claimed in claim 1 further comprising a workbench module coupled to the processor, the workbench module is configured to generate reports having the claims data.

10. The system as claimed in claim 9, wherein the workbench module is configured to illustrate the one or more suspicion indicators in the plurality of claims upon receiving a drill down request from a fraud analyst.

11. The system of claim 1, wherein the fraud dictionary is updated over a period of time by adding new concepts, wherein the new concepts are generated by means of machine learning techniques to learn new patterns from the at least one of insurance claims fraud and subrogation potential claims so identified.

12. A method for identifying insurance claims fraud and subrogation potential claims in an insurance industry, the method comprising:

extracting, by a processor, claims data from one or more data sources, wherein the claims data includes structured data and unstructured data, and wherein the unstructured data includes imaged content and speech data;

converting, by the processor, the imaged content and the speech data to a textual format, wherein the imaged content includes photos and scanned copies of claim related information;

combining, by the processor, the claims data, having the textual format of the imaged content and the speech data, to obtain a consolidated claims record having text data, wherein data, from the claims data, in a language different from a language used in most of the claims data is translated to the language used in most of the claims data;

removing, by the processor, noise from the text data of the consolidated claims records to obtain a claim dataset, wherein the claim dataset includes a plurality of claims;

ascertaining, by the processor, one or more suspicion indicators in the plurality of claims based on concepts obtained from a fraud dictionary using one or more analytical technique, wherein the concepts are at least one of words, phrases, and entities that indicate suspicious claims;

assigning, by the processor, a score to each of the plurality of claims based on at least one scoring rule comprising one or more scoring parameters, wherein the score is an indicative of a level of suspicion of a claim; and

identifying, by the processor, at least one of insurance claims fraud and subrogation potential claims based on the score assigned to each of the plurality of claims.

13. The method as claimed in claim 12, wherein the one or more data sources include at least one of internal and external data storage systems in an insurance organization, the imaged content which contains the photos and the scanned copies of the claim related information, the speech content, and web content.

14. The method as claimed in claim 12, wherein the one or more analytical technique is at least one of a text analysis, a predictive model analysis, a link analysis, business rules analysis, and a social-media analysis, and wherein the text analysis uses at least one text pattern to identify the suspicion indicators, and wherein the link analysis uses suspicious links between entities of two different claims to identify the suspicion indicators, and wherein the business rules analysis uses business rules to identify the suspicion indicators, and wherein the social-media analysis uses social-media content and social-media links between entities involved in a claim, from the among the plurality of claims, to identify the suspicion indicators in the claim.

15. The method as claimed in claim 14, wherein the fraud dictionary includes the at least one text pattern, the business rules, rules for identifying the suspicious links between entities of two different claims, and rules for identifying the social-media content and the social-media links.

16. The method as claimed in claim 12, wherein the one or more scoring parameters are a combination of at least one of rule name, client, line of business (LOB), suspicion indicators, claim filtration criteria, and claims funnel.

17. A non-transitory computer-readable medium having embodied thereon a computer program executable in a computing device for executing a method of identifying insurance claims fraud and subrogation potential claims, the method comprising:

extracting claims data from one or more data sources, wherein the claims data includes structured data and unstructured data, and wherein the unstructured data includes imaged content and speech data;

converting the imaged content and the speech data to a textual format, wherein the imaged content includes photos and scanned copies of claim related information;

combining the claims data, having the textual format of the imaged content and the speech data, to obtain a consolidated claims record having text data, wherein data, from the claims data, in a language different from a language used in most of the claims data is translated to the language used in most of the claims data;

removing noise from the text data of the consolidated claims records to obtain a claim dataset, wherein the claim dataset includes a plurality of claims;

ascertaining one or more suspicion indicators in the plurality of claims based on concepts obtained from a fraud dictionary using one or more analytical technique, wherein the concepts are at least one of words, phrases, and entities that indicate suspicious claims;

assigning a score to each of the plurality of claims based on at least one scoring rule comprising one or more scoring parameters, wherein the score is an indicative of a level of suspicion of a claim; and

identifying at least one of insurance claims fraud and subrogation potential claims based on the score assigned to each of the plurality of claims.

18. The non-transitory computer readable medium as claimed in claim 17, wherein the one or more analytical technique comprises a text analysis, a predictive model analysis, a link analysis, business rules analysis, and a social-media analysis, and wherein the text analysis uses at least one text pattern to identify the suspicion indicators, and wherein the link analysis uses suspicious links between entities of two different claims to identify the suspicion indicators, and wherein the business rules analysis uses business rules to identify the suspicion indicators, and wherein the social-media analysis uses social-media content and social-media links between entities involved in a claim, from the among the plurality of claims, to identify the suspicion indicators in the claim.

19. The non-transitory computer readable medium as claimed in claim 17, wherein the one or more scoring parameters are a combination of at least one of rule name, client, line of business (LOB), suspicion indicators, claim filtration criteria, and claims funnel.

20. The non-transitory computer readable medium as claimed in claim 17, wherein the plurality of claims are categorized into one or more categories based on the score assigned to each of the plurality of claims.