www.niit-tech.com Variable Auto-Adjudication Changing the Paradigm NIIT Technologies White Paper
CONTENTS Today s Challenge 3 The Fraud Management Solution Variable Auto-Adjudication 4 How It Works 4 Data Noise Reduction 6
Auto-Adjudication Today Insurers use a standard claim management process called auto-adjudication to process and pay a claim amount if it is less than a preset threshold amount. The threshold amount is calculated and fixed by the insurers on the basis of the line of business and other factors. It is kept low to ensure that a high volume of claims go through the auto-adjudication process with minimal random audit. The number of claims that are automatically adjudicated by insurers is increasing. As an example, 79% of claims in healthcare were adjudicated automatically in 2011, as compared to 37% in 2002. 71% 68% 49% 44% 37% 27% 2002 2006 Paper 80% 75% 80% 79% 53% 37% 2009 2011 Electronic All Sample Claims Fraud (Texas Department of, 2010 annual report) Worker s Compensation 10% Title/Escrow 16% Property & Casualty 6% *Other 3% Commercial 5% Motor Vehicle 30% Credit Life/ Disability 11% Health 6% Homeowners 11% Life/Annuity 2% The insurance industry consists of more than 7,000 companies that collect over USD 1 trillion in premiums each year 1. The massive size of the industry contributes significantly to the cost of insurance fraud by providing more opportunities and bigger incentives for committing illegal activities. The total cost of the U.S. insurance fraud (non-health insurance) is estimated to be more than USD 40 billion per year and USD 80 billion overall (Coalition against Fraud). That means the insurance fraud costs an average American family between USD 400 and USD 700 per year in the form of increased premiums for non-health insurance. Source: AHIP Center for Policy and Research. There seems to be a much greater threat in the insurance vertical with the proliferation of transnationals, also known as international bad actors. Transnationals try to find information from both large and mid-market insurers by hacking, cyber probing and finding an insider who could disclose the threshold claim limit. An illustration of the breakdown of the fraud claims reported by the Texas Department of in their 2010 annual report identifies estimated percentages of fraud that are understated as compared to the actual extent of fraud. In late August 2005, Hurricane Katrina made landfall along America s Gulf Coast. It caused widespread destruction and cost approximately USD 100 billion in terms of economic damages. Approximately 1.6 million insurance claims were filed, totaling USD 34.4 billion in insured losses. Of the USD 80 billion in government funding appropriated for reconstruction, it is estimated that insurance fraud may have accounted for as much as USD 6 billion. 2 Auto-adjudication managed by artificial thresholds does not stop fraud because the process is designed to make claim handling more efficient and cost-effective. 1. http://www.fbi.gov/stats-services/publications/insurance-fraud 2. http://www.fbi.gov/stats-services/publications/insurance-fraud 3
The Fraud Management Solution Variable Auto-Adjudication By injecting a real-time variable auto-adjudication process into claims handling, the insurer can continuously identify and prevent fraudulent claims from being paid or settled before preventive actions are taken. The variable auto-adjudication solution from NIIT Technologies identifies claim fraud in real-time above and below the standard thresholds that an insurer chooses, thus changing the paradigm. The variable process utilizes predictive analytics to determine the relationships of domestic and transnational bad actors preventing them from gaming the insurer s processes. NIIT Technologies Digital Foresight a predictive analytics solution has the potential to reduce fraud on auto-adjudication claim processes in a significant percentage that can save insurers up to USD 400 million annually. Reduces SIU Costs How It Works The current standard practice has insurers evaluating risks from experience and using their knowledge stack of pre-defined rules to set the claim threshold amount for the auto-adjudication process. Digital Foresight s Variable Auto-Adjudication plans to go beyond the standard practice of using predictive analytics to prepare for the next generation of risk management and mitigation. This complex data-science-driven, machine-learning-enabled cognitive computing engine incorporates a broad range of factors including geospatial intelligence factors to dynamically recalibrate the threshold claim amount on a case-by-case basis. The factors are weighted differently depending on the application parameters - making fraud impossible in the claims auto-adjudication process. Auto-Adjudication Claims below the threshold paid according to the operating, financial and regulatory framework. Failure is realized due to... Designed for financial and operating efficiency in people, process and service Set dollar threshold QC runs with periodic operating and financial audits Reduces Loss Ratio Payment Distr. Notice Programs Claims Mgt. Real-time Claims Mgt. Auto-Adjudication cannot detect adverse patterns early in the process and works best when bad actors come in play. Beyond financial losses, the data blurred under auto-adjudication dumbs the insurer further. Applies to every bill under the claim Compliance Oversight Direct Contract Mgt. Settlement Admin. Call Center Strips out fraud-based slips under conventional claims systems Database Mgt. Digital Foresight Benefits Identifies fraud on first instance Variable Auto-Adjudication Uses an AI-driven variable payment threshold; Variables reset from a +100 factor stack on random days, time, amount, location, vendor, etc. configurations. Variable Auto-Adjudication detects adverse patterns early in the process and defeats even advanced gaming used by bad actors. Reduced Loss 1. Designed for financial and operating efficiency in people, process and service 2. Variable dollar threshold 3. QC runs concurrently and with periodic operating and financial audits The financial losses caught and reduced through deep, real-time data points - enables insurers to be proactive in customer, product, and market distribution and financial management. 4
This machine-driven solution also eliminates the internal points of compromise as they cannot disclose the machine-driven limits to bad actors. Digital Foresight s Variable Auto-Adjudication solution applies not just to a claim but to every bill under the claim. Digital Foresight s Claims Actionable Intelligence, provides a better understanding of the entire claim ecosystem where people, behavior patterns, places, things, locations and time can be projected. People Objects O6 O4 O6 O5 O5 O5 Structures Locations S3 S1 S4 S2 Events/Time Behavioral Events/Time Trajectories State Space Models of Entities Agent - Based Modelling A B D F G H C E I Events/Time Variable Auto-Adjudication recognizes that bad actors behavior is identifiable using the most pertinent and relevant data from internal and external sources in real-time. Internal claim data (historical and current) provides relevant open domain source information, data and intelligence fused with internal data when needed. Surface Net Deep Net Dark Net Getting to all relevant information: Surface Net: 20 TB, 1Bn unique records, 100% public access Deep Net: 6,500 TB, 550Bn unique records, 95% public access Claim Counter Fraud Solutions from Real People with Real Experience Data science Geospatial experts All source intelligence experts Experts in web IT Intelligence systems experts Best-in-class tools Real insurance experience: - Claims - Underwriting - SIU - Finance & Actuarial - Regulatory & Rating - Distribution 5
Data Noise Reduction Noisy cells in the data are prioritized and ranked in their critical value Digital Foresight supports variable auto-adjudication by managing data noise. Change is constant in Digital Foresight, and so is data noise. Data warehouses and repositories are known for making constant noise. So the confidence in the output and the use of a based on the output needed or desired. The Next Best Action is to determine what external sources of information and intelligence can be reached to provide contemporaneous (and credible) data that can be incorporated in the internal data repository. number of models results in a failure to achieve objectives and ROI. The quality of the data generated is not adequate to provide Digital Foresight has the ability to reach all viable, applicable, and valuable customer information. When you step back and consider relevant information from external sources not just restricted to the robust nature of the source of information and intelligence social media. In short, Digital Foresight enables internal data noise harvested from digital intelligence, there is a ready solution to the to be cured and maintained to provide the best information for noisy data problem. insurers to combat claim fraud. Internal data repositories Digital Foresight harvests verified and intrinsically have noise - reducing validated data from external information accuracy and value. and intelligence to align it with prioritized data free from noise. Digital Foresight maps data for weakness and noise. It also prioritizes data needed to enhance and optimize it for usage in the customer s specific mission space. Open Source Intelligence Geospatial Intelligence Continuous flow of verified and Social Media validated information and Non-Social Media Contextual Information intelligence in the data repository. The result is the customer realization of the next best action based on optimized data. Digital Foresight s proprietary tradecraft, best-in-class tools and expertise develop the right, verified and validated information needed to meet the customer mission space. Includes key expertise in applying signature analytics and behavior trajectory to machine learning protocols to enable enhanced data to flow freely to the customer mission space. Digital Foresight : Real-time output to Know Tomorrow, Today 6
About NIIT Technologies NIIT Technologies is a leading IT solutions organization, servicing customers in North America, Europe, Asia and Australia. It offers services in Application Development and Maintenance, Enterprise Solutions including Managed Services and Business Process Outsourcing to organizations in the Financial Services, Travel & Transportation, Manufacturing/Distribution, and Government sectors. Employing over 8,000+ professionals, NIIT Technologies follows global standards of software development processes. Over the years the company has forged extremely rewarding relationships with global majors, a testimony to mutual commitment and its ability to retain marquee clients, drawing repeat business from them. NIIT Technologies has been able to scale its interactions with marquee clients in the BFSI sector, in Travel Transport & Logistics, and Manufacturing & Distribution sectors, into extremely meaningful, multi-year collaborations. NIIT Technologies follows global standards of development, which include ISO 9001:2000 Certification, assessment at Level 5 for SEI-CMMi version 1.2 and ISO 27001 information security management certification. Its data center operations are assessed at the international ISO 20000 IT management standards. India NIIT Technologies Ltd. Corporate Heights (Tapasya) Plot No. 5, EFGH, Sector 126 Noida-Greater Noida Expressway Noida 201301, U.P., India Ph: + 91 120 7119100 Fax: + 91 120 7119150 Americas NIIT Technologies Inc., 1050 Crown Pointe Parkway 5 th Floor, Atlanta, GA 30338, USA Ph: +1 770 551 9494 Toll Free: +1 888 454 NIIT Fax: +1 770 551 9229 Europe NIIT Technologies Limited 2 nd Floor, 47 Mark Lane London - EC3R 7QQ, U.K. Ph: +44 20 70020700 Fax: +44 20 70020701 Singapore NIIT Technologies Pte. Limited 31 Kaki Bukit Road 3 #05-13 Techlink Singapore 417818 Ph: +65 68488300 Fax: +65 68488322 A leading IT solutions organization 21 locations and 16 countries 8000+ professionals Level 5 of SEI-CMMi, ver1.2 ISO 27001 certified Level 5 of People CMM Framework Write to us at marketing@niit-tech.com www.niit-tech.com D_96_280515