HIGHLIGHTS OF THE 2023 AHOU CONFERENCE – WAVES OF CHANGE: NAVIGATING THE UNDERWRITING EVOLUTION TOGETHER
Bonnie Stroder
OTR Editor-in-Chief bonnie.stroder@rgare.com
Tami Britton
OTR Executive Editor britton.tamarah@pennmutual.com
Anne Bunkers
OTR Administrative Editor anne.bunkers@prudential.com
The 22nd annual AHOU Conference was held in Hollywood, FL, April 16-20, with 982 attendees. The conference opened with AHOU President Jennifer Johnson welcoming all attendees. AHOU Past President Carolyn Goshorn presented this year’s Hall of Fame inductees William Tilford and Elizabeth Collier. Jodi McDonald, ALU President, introduced the new FALUs in attendance and presented them with diplomas.
Attendees also learned of the many ways ALU has grown and how everyone can all get involved. ALU President Jodi McDonald discussed ALU webinars, the Certified Education program now being offered for FALUs and underwriters, and exam results and statistics. A special shout-out goes to the 79 new FALUs this year. The Charles A. Will award winners for 2020 and 2022 were announced. The Charles A. Will award has been given annually since 1993 to the newest Fellow of the Academy of Life Underwriting (FALU) who obtains the highest aggregate score across all four ALU exams. The award is named for the former Cologne Re (now GEN Re) senior underwriting executive, Charles Will, who is known for his 1974 book, Does it Make Sense? The 2020 Charles A. Will award winner was Kevin Alameddine, Swiss Re Canada, who was unable to attend AHOU until this year. The 2022 winner was Sheetal Salgaonkar, RGA Global, India.
New to the AHOU was a recently developed mentoring program. The launch of the program began with first-time attendees being matched up with experienced AHOU attendee members. The mentors greeted the first-time attendees on Sunday to show them around the Solution Showcase and answer any questions they may have had, and continued to build a relationship throughout the conference. The official AHOU mentoring program launched even further with an opportunity for all AHOU members to sign up to be a mentor, mentee or both.
The conference offered many tracks for attendees to choose from during each session time period: the Core track, Industry Insights track, Career Development track, Innovation track and Risk Management track. Many concurrent sessions were repeated at different times for attendees who wanted to attend multiple sessions that were offered at the same time.
The mainstage presentation after the opening ceremonies was Shark Tank, hosted by Chris Behling of Northwestern Mutual.
Shark Tank: Whose Vision for the Future of UW Will Prevail?
Chris Behling, Vice President, Underwriting, North-Western Mutual
Norm Leblond, Vice President, Chief Underwriter and Claims Risk Officer, Sun Life Canada
Lauren Cross, Assistant Vice President and Actuary, John Hancock
The presenters were tasked with pitching their ideas for the future of underwriting to three judges. Sears Merritt was unable to attend, so Chris Behling filled in as a presenter along with being the host. Judges were Nichole Myers, Chief Underwriter of Ethos; Catie Muccigrosso, RGA; and Jessica Caracofe, MIB.
AHOU Executive Council (left to right) front row: Whitney Barnes, Kris Chaney, Carolyn Goshorn, Jennifer Johnson, Mark Lounds, Jessica Caracofe; back row: Catie Muccigrosso, Jonathan Meindel, Michelle Privett, Ken Turscak, Kim Waterfield, Mary Hanson.
From the April AHOU Conference in Hollywood, Florida: Outgoing AHOU President Jennifer Johnson passes the gavel to incoming President Mark Lounds.
Norm Leblond presented an idea to develop a solution to underwriting as created by underwriters for underwriters, challenging the way we look at underwriting. The idea was innovation created to leverage artificial intelligence (AI), and was titled Therapeutic Interviews. This would leverage ChatGPT and would ping digital information at the start of the application, feed the data models in the background that would create questions based on the information received from the rules engine and models already in place. This would create a dynamic application, and become more of a conversation rather than an application.
Chris Behling introduced a concept called U.M.A.P. (Universal Mortality Assessment Program), which is a universal underwriting program. This program would create a risk score in a way similar to credit scores and would be applied similarly. The model would look at all the data obtained, and create a universal score that would be used to determine a person’s insurability and risk class. The model would also provide information on what a person would need to do to improve their mortality score.
Lauren Cross presented Healthy Habits for Life. This would be an offering for a 3-year term product that would be easy to apply for and fast approval using a health questionnaire and third party data. While 52% of US citizens have life insurance, there is concern for it not being affordable, especially to those with chronic health conditions, blue-collar workers, stayat-home moms, etc. Healthy Habits for Life would provide some coverage for everyone at an affordable cost. The basis of this is to utilize a smartphone app that would allow people to log healthy habits to accumulate credits to extend their coverage longer than the 3-year period.
The judges interacted with each “shark” with questions on how their ideas would be utilized and the impact they felt it would have on applicants for coverage. At the end of the presentation, an interactive poll of the audience determined the winner.
Data for Good: Data Bias Governance and the Regulatory Landscape
Christie Corado, Esquire, General Counsel, Privacy Officer and Corporate Secretary, MIB, LLC
Tom Fletcher, PhD, Vice President, Data Analytics North America, PartnerRe
Scott Kosnoff, Partner and Co-Chair of the firm’s Artificial Intelligence, Algorithmic Decision-Making and Big Data Team, Faegre Drinker Biddle & Reath Patrick Sugent, Vice President, Data Science, Lexis-Nexis Risk Solutions
This was a presentation regarding data bias with the tools used to help streamline the underwriting process. The discussion covered quantitative testing for unfair bias and the sources that are established to test for unfair bias; data responsibility and how regulators want to see transparency; how the data is used and that the data has been verified. The presentation also covered the pros and cons of ChatGPT and how this information is easily accessed. However, not all the information is correct and using this information should be done with caution. There was also a discussion on the expectation of third-party vendors and the governance of the information they provide.
Liar, Liar, Pants on Fire: The Art of Live Rescissions
Amy Rider, FALU, FLMI, Vice President, Underwriting Strategy and Mortality Management, Sammons Financial Group
Susan Mayer, FALU, FLMI, Vice President, Underwriting Specialty Products, Hannover Re
This interactive session with case studies focused on reviewing the process of living rescissions and material misrepresentation on applications.
Material misrepresentation is an untrue statement on an application form that would affect the underwriting decision. Material misrepresentation on applications costs the average customer 5-10% in higher premiums. Overall, life insurance fraud costs the industry $74.7 billion annually.
Best practices on assessing material misrepresentation:
1. Does the misrepresentation found on the application align with an application question?
2. Does the application question allow for the history to be disclosed?
3. Does the misrepresentation align with the timeline of the application question?
4.Was the applicant knowledgeable about the condition or test results?
5. Would the company consider this misrepresentation as material?
6. Does the company’s written underwriting practices position it favorably if the rescission goes to litigation?
7. Are there extenuating circumstances?
A live rescission is the process by which a life insurance policy may be voided during the contestable period due to material misrepresentation on the application while the insured is alive. A live rescission takes an average of 2 hours to process vs. the time it takes to review a contestable claim, which can take weeks to investigate and process.
Fundamentals of a live rescissions:
1. May reduce early duration claims.
2. Reduce fraud due to the sentinel effect.
3. Reduce the time needed to review fraudulent claims.
4. Reduce negative mortality impacts to the business.
Each company has its own set of criteria used to determine whether to rescind a policy. There is a high percentage of companies who automatically rescind policies if the misrepresentation would have resulted in a decline, had the information been disclosed, or if there is tobacco use identified and the policy was issued at non-tobacco rates. While a majority of companies use the threshold of two tables or two risk class differences, some companies will allow three or four table/class differences prior to sending a policy for rescission.
The opportunity for misrepresentation exists in all lines of business.
• Traditional underwriting -> Opportunity exists for misrepresentation.
• Accelerated underwriting -> Higher opportunity for misrepresentation.
• Direct to consumer -> Highest opportunity for misrepresentation.
Best practices for investigating material misrepresentation:
1. Use multiple underwriting tools.
2. Have written policies and procedures.
3. Review state-specific regulations.
4. Perform trend analysis and make recommendations for process improvements.
5. Enhance authorizations.
Underwriting tools available for investigating material misrepresentation:
1. MIB Plan F.
2. Prescription drug data.
3. Attending physician statement (APS).
4. Electronic health records (EHR).
5. Medical billing data.
6. Aggregated lab data.
7. Criminal records.
Each tool has its own pros and cons. Depending on the misrepresentation, using one or more of the available data should be considered.
It is important to make sure there is documentation regarding the policy and procedure the company follows to process a living rescission. Many disciplines may need to be involved, especially working with the legal department on each rescission to make sure the rescission is in line with state-specific regulations.
Companies should perform trend analysis on the findings for living rescissions. Reviewing the reason a policy is rescinding can help identify areas where a question may need to be reflexed or reworded to capture information that may not currently be clear on the application. Underwriters performing postissue reviews and contestable claim reviews are having more of a voice in helping with question wording and reflex questions.
One of the challenges of obtaining post-issue records is the special authorization required by some physicians or medical facilities. Review the authorization signed at the time of the application and, if necessary, add a section to the authorization that will allow your company to retrieve the records when a special authorization is required. Hannover Re is currently piloting this “special authorization” form with some of its clients.
Living rescissions need to be objective and specific and not based on opinion. A company must be able to defend the case in a court of law. If the case cannot be defended, the policy should not be rescinded.
Best practices on letters to owners:
1. Company relies on the truthfulness of application answers.
2. Details and source of the application answer believed to be false along with a copy of the application.
3. Underwriting decision if the application answer was truthful.
4. Opportunity to provide information to refute.
5. Explanation of a live rescission, including return of premium.
6. Choices of options and implications.
7. Signature section with statement of indemnification to the company against future claims.
8.Timeline on next steps needed.
Best practices on letters to financial professionals/ agents:
1. Company relies on the truthfulness of application answers.
2. Details and source of the application answer believed to be false along with a copy of the application.
3. Underwriting decision if the application answer was truthful.
4. Impact to commissions.
Liquid Biopsies – The Next Big Thing in Medicine ? or !
Dr. Achim Regenauer, MD, Chief Medical Officer, Partner Re
The incidence of cancer in ages under 50 is increasing. Cancer screening is important in detecting cancer early.
Cancer screenings are available for some cancers:
• Breast cancer – Mammograms/ultrasounds.
• Colon cancer – Colonoscopies, Cologuard.
• Lung cancer – Chest X-ray (CXR) and CT scans.
• Cervical cancer – Pap smear.
• Prostate cancer – PSA, prostate MRI, digital rectal exam (DRE).
Diagnosis of cancer, as well as stage of cancer at diagnosis, matters. When diagnosing cancer, early stage/ early detection is key in treatment. Advanced stage is considered Stage 3 and Stage 4. Distant metastasis is found in 170 out of 100,000 cases. In the US, 71% of cancers are not found early.
Screening is key to finding cancers early. For example, the incidence of cervical cancer shows that mortality is going down. The incidence of thyroid cancer is increasing; however, the mortality is not changing. There is a lack of established screening protocols for many cancers, which is why screening tests and the prevalence of testing are low. There is a medical lack of knowledge of early stage of some cancers and there is no knowledge of tumor behavior at early stages.
The dynamic of the cancer is also important. Cancer has different growth speeds – fast, slow, very slow and non-progressive. Acute lymphocytic leukemia (ALL) is a fast-growing cancer. Lung cancer is a slow-growing cancer. Colorectal cancer is a very slow-growing cancer. Thyroid cancer is a non-progressive cancer.
The gold standard for determining cancer is a tissue biopsy. However, the tissue biopsy is invasive, painful and time-intensive. With tissue biopsies, there is confirmed diagnosis when underwriting a policy and low risk for adverse selection.
Liquid biopsy is the future and all about timing. Liquid biopsies can be performed on any bodily fluid – blood, sputum, urine. The test must be performed at different intervals as people age. The test may be negative for cancer at age 50, but at age 55 there may be a positive screening for cancer. The test will assist in early detection of cancer, is cheaper to perform, and uses molecular profiling. With liquid biopsies and underwriting, there is a higher risk of adverse selection as the diagnosis of cancer is not definitively confirmed.
The most notable liquid biopsy is GRAIL Galleri testing. This is an AI-enabled evaluation that screens for 50+ different cancers. The test does not require a physician prescription and takes about 10 days to complete. However, the cost of the test is $950 and is not reimbursed by insurance.
There are both harms and benefits of liquid biopsies, and, at present, more questions than answers.
Clinical practice:
• Definition of target group(s)?
◦ Age
◦ High-risk groups
• Frequency/interval of testing
• Which cancers should be targeted?
• Many MCED tests
◦ Which one?
◦ Lack of standardization
• Not a diagnostic test
• MCED test negative – Is it really negative? Follow-up?
• MCED test positive – Workflow, diagnostic workup? Therapy, watchful waiting, pre-emptive, etc.?
Insurance:
• Not a prediction such as BRCA test, but a snapshot moment only.
• Galleri not the only player, but the most promising one.
• No framework yet–insurance reimbursement and regulation are key.
Game-changer potential – breakthrough and clinical application in this decade likely.
Scams & Schemes: How Underwriting Can Detect Fraud
Ken Elder, CPA, FLMI, Vice President, Enterprise AML Officer, Lincoln Financial Group
Kevin Glasgow, ARA, FLMI, FLHC, CLU, CFE, Vice President, Investigation Solutions, Diligence International Group, LLC
James Swinton, FLMI, FALU, Second Vice President, Munich Re
There are many different types of fraud that occur in the life insurance industry. Being aware of the different methods of fraud is key in identifying fraud and preventing excess mortality for your company.
Producer fraud:
• Rebating and advanced commission schemes.
◦ Fast pace of application submission with the same product.
◦ Premiums consistently matching target premium.
◦ High rate of early lapse.
Policy situs fraud:
• Application is taken in one state but solicitation in another state.
• Exams.
• Sales communications -NY, Washington State have strict insurance laws.
Conflicts of interest:
• Trustee
• Owner
• Beneficiary
• POA
There are multiple trends in committing fraud. Some of these trends are opportunistic fraud, reactionary fraud, systemic fraud and synthetic identity fraud.
Opportunistic fraud/reactionary:
• Medical misrepresentation on the application.
• Wagering contracts.
Systemic fraud:
• Wagering contracts.
• Imposers posing as insured: stolen IDs, synthetic identities.
Synthetic identity is a fake identity that combines real personal information with fraudulent identity. The fraudster will use the fake identity multiple times to establish the identity with the credit bureau. For example, the fraudster will apply for a credit card and get denied. They will apply again and again and again until they are finally approved and credit is established. This will then create the synthetic identity as a true identity.
Some fraud is not found until a claim is submitted. At claim time, it is found that there is no insurable interest between the insured and beneficiary. There are also incidences of switched identifies (real or created) and even switched bodies. It is also found that the insured had multiple identities. In some of these incidences, it was found that the funeral home was involved in perpetrating the fraud. Death certificates do not match hospital records.
So how do you handle it when fraud is found during the claim investigation and you are outside of the 2-year contestable period? Most companies do not realize that it is still possible to void the policy outside of this contestable period. The contestable period would not apply, as the contract was not valid due to the fraud that was perpetrated.
Training staff to identify fraud can mitigate putting fraudulent policies in force. There are many ways to identify red flags for possible fraud, including verifying addresses, reviewing emails provided, reviewing the insurable interest of the beneficiary, questioning owners if the proposed insured is an adult and not owning their own policy.
Mitigating fraud:
• Train staff on fraud risks
• Stay current on trends
• Review early claims
• Feedback and communication
Mitigates to wagering:
• Know the identity of the applicant.
• Know the location of the applicant and application.
• Monitor contract information.
• Review all early deaths.
• Question insurable interest.
• Question limited digital footprints.
• Question applicant if something sounds inconsistent.
• Fight fraud when found. Remedies may exist outside the 2-year contestable period.
Fraud prevention tools:
• MIB – jumbo risk tool
• ID verification
• Lab reports
• Public records
• MVRs
• Prescription drug profiles
• Other data sources
Jumbo risk tool – MIB:
• Provides inforce information
• Provides information on lapsed policies
• Provides patterns of policies applied for
Why location matters:
• VPNs can fool IP addresses, making it difficult to determine location.
• Geofencing programs – provides GPS coordinates and can give a definite address.
• High-risk locations can be identified.
Material misrepresentation on the application:
• Cancer is most impactful impairment.
• For accelerated UW, extremely important to make sure the application question wording gives you a solid opportunity to defend the misrepresentation (Ex: alcohol).
• Underwriting needs to collaborate with the claims department on trends.
• Telemetry data models can assist in drafting effective questions (change from “have you ever...” to “how many times have you…”).
• Non-medical questions – the more detailed the better.
• Direct-to-consumer (DTC) products have a significant higher rate of criminal history.
• MVRs – some states do not have a reciprocal state program to report violations in states other than where the insured is licensed.
• Foreign risk misrepresentation - knowing your financial professional/agent is very helpful. Most foreign risks have clean-sheeted applications from a medical standpoint.
Mitigating Misrepresentation
Marlee Nichols, FSA, Assistant Director, Risk Selection Analytics, Northwestern Mutual
Many of the presentations are discussing the post-issue process and investigating fraud and material misrepresentation. But we can head off issuing policies prior to approval to prevent material misrepresentation and fraud by promoting honesty of the applicants when completing their application for coverage.
Why is it important to promote honesty to applicants?
• $300 billion in fraud, $75 million in life insurance alone.
• 2-year contestability.
• Evolution of accelerated UW has increased the risk of misrepresentation.
CRL reported $4 billion in tobacco cost to insurance industry by misrepresentation of tobacco use on the application.
Misrepresentation can take years and years to surface as to the cost of the misrepresentation to the company. The cost surfaces in later years, not during the contestability period.
An exam of misrepresentation of build by 70 pounds can cost a company $14,000 over the duration of the policy. If there is misrepresentation of both build and blood pressure, the cost to the company can double. The cost of tobacco misrepresentation can total $40,000 or more in premiums.
Tools to promote truthfulness in applicants:
• In DTC: Behavioral economics, digital data and modeling concepts.
• In intermediated distribution: Same as DTC along with distribution monitoring.
Behavioral economics:
• Anchoring techniques: Input height and then provide a range for average weight for people at that height along with factoring in a few pounds higher to promote truthfulness.
• Have the applicant sign an honesty attestation.
• Choice architecture: How are we asking the questions? The way the questions are being asked can incentivize people to answer in a certain way.
• Supervision: People disclose information differently when supervised vs. unsupervised selfdisclosure online.
• Hawthorn vs. Sentinel Effect.
◦ Hawthorn: Knowing someone is watching or observing.
◦ Sentinel: Follow-up and secondary consequence. Letting people know upfront that information will be obtained to verify their answers.
Distribution monitoring:
• Monitor risks across distribution channels.
• Improving partnerships.
• Perspective from the field: field underwriting.
Tools: Digital data:
• Where are applicants signing an authorization?
• Benefits of digitally available data.
◦ Mortality savings.
◦ Improved client experience.
• What new sources are available to obtain the same data we obtained in traditional underwriting?
Modeling concepts:
• Can what we already know be used to inform new modeling concepts?
• Models targeted at finding honest disclosure.
Regulatory and ethical concerns:
• Ethical and legal concerns from focusing on liars.
• Certain variables may lead to unfair bias.
• Important to understand potential sources of unfair bias.
Precision Medicine, Early Disease Detection and Treatment: 21st Century Approaches to Using Genetics and Epigenetics
Wayne Lilyestrom, Medical Director, GRAIL LLC Glenn Maston, PhD, Senior Staff Genomic Variant Scientist, ExamOne
Andrea Paal, MS, LCGC, Genomic Science Liaison, ExamOne
Rob Philibert, MD, PhD, CEO, Behavioral Diagnostics
Precision medicine is the idea that personal data, especially one’s genetic data, can be used to tailor a person’s individual treatment. Our DNA is our recipe for life, and everyone’s DNA carries some variants. Proteins are the molecular components of our DNA and each one carries out a specific task. If these proteins cannot perform their job properly, it may cause disease. While some of these variants do not alter the function of the protein, others do. Precision medicine depends on genetic testing and the correct interpretation of the variants.
Past medical treatment was based on a one-size-fitsall structure. You have a disease/condition, you receive the same treatment. Present medical treatment is personalized to our unique genetic profile. The goal of precision medicine and personalized medical treatment is to avoid adverse reactions, reduce trial and error, and minimize time and money. Future medical treatment will be based on gene editing. There is exciting new technology called CRISPR-Cas9 which provides correction of the variant, thus allowing us to edit the DNA variant and to remove the source of disease. The first clinical trial in humans has been approved and completed by the FDA and is on sickle cell disease.
Epigenomes are secondary modifications that affect the expression of the DNA code. DNA methylations are the on-and-off signs for the genome. There is a “reference” genome that is the most common sequence. We can compare a person’s genome to the reference and variants can cause disease. Proteins play a part in this genome sequence. Pathogenic variants in proteins may disrupt function, reduce function or even increase function.
Epigenetic effects are large and are acquired through the environment. Environmental factors such as smoking and alcohol consumption play a big role in the variants seen. Environmental factors and not genetic factors are main determinants of most common disease. The testing today identifies variants but cannot tell if this variant causes disease.
The economic toll from cancer is staggering. Years of lost life = $8.7M. Annual cost of cancer = $207 billion. Economic burden = $1.3 trillion which includes care givers. Tumors shed nucleic acids into blood, carrying cancer-specific information. The more aggressive or size of the tumor, they usually shed more. DNA sequences can on occasion signal in the blood before showing up on any scans. However, not all cancers shed signals into the bloodstream. There are many studies in place, with the use of AI, to detect cancer signals.
Key performance of these tests:
• Over 50 cancers detected.
• False positive rate of 0.5%.
• When detected, 89% of the time, find the location of the tumor/cancer.
So why not offer this testing to everyone? It is costprohibitive; a single test right now is approximately $1,000, and not covered by insurance. Also, how often should you have this testing done since as we age, we accumulate damage to our genome. We also need to think about disparity in care; if you are diagnosed with an early-stage cancer, do you have the resources to treat it? However, think of the potential of removing cancer from the population stage by stage. We could catch the early stages and they may never reach the higher stages that tend to be harder to treat and usually more costly.
In the underwriting arena, this could be a gamechanger. If we could offer or see the results of this testing, it could change our decisions exponentially. Also, if the population is detecting cancer at an earlier stage and treating it, then we could possibly offer more quickly.
It’s Never Too Late To Learn: Ongoing Training for the Seasoned UW
Mark Dion, FALU, FLMI, VP, Underwriting Education & Training, RGA
The first rule of underwriting is: Everyone dies. The second rule of underwriting is: We don’t know when.
Look at your underwriting career in stages. There are four stages of an underwriting career:
1. The Jet Underwriter, or the first 2 years. You are learning the basics of insurance principles, body systems, non-medical risks, etc.
2. The Consultant, 2-5 years of experience. You are continuing to underwrite in a production environment, strengthen relationships and use your critical thinking skills.
3. Sr. Consultant, Leadership, Technical, Specialty, 5-10 years of experience. You are developing your underwriting skills, project management skills and leadership skills. This is the stage where a lot of training should be. Figure out your strengths, gaps and potentials.
4. Sr. Professional, Management, 10 years and up. Provide training and leadership, presentations, technical expertise. Use your skills to train others.
We, as underwriters, have many career paths. Remember, everyone doesn’t get to be chief. We can become subject matter experts, trainers, a claims liaison, marketing underwriter and team leaders. Think of things outside of the norm and realize there are many avenues we can pursue.
Continuing education is very important in an underwriting career. Once you get your FMLI or FALU, it doesn’t stop there. Take a college course, how about claims investigation, get your continuing education credits, take an EKG reading course. There are many reasons to continue your education including:
• Increase chances for promotion.
• Improve your image and marketability.
• Increase your salary.
• Increase personal development.
• Increase your ability to make career changes. The more skills you have, the more marketable you become. Remember, every piece of paper means something.
We must also develop a lifelong learning model for ourselves and never stop learning. The world does not stand still and there are always evolving medical breakthroughs, changes in treatment, AI, etc. Be sure to keep up with the research and learning. When there is a fork in the road, take it! Sometimes our road turns in unexpected ways. Embrace change. Also, be sure to be a good mentor and leave a legacy. Teaching and mentoring are the best way for us to learn.
AI in Insurance – ABehind-the-Scenes Look
Lovell Hodge, PhD, VP, Data and Adaptive Intelligence, Munich Re
Artificial intelligence (AI) is “the theory and development of computer systems that learn, understand and deal with new situations in a seemingly intelligent way.” Traditional programming solves the problem, but only the problem it was built to solve and in the way it was designed to solve it. This is a different way of solving problems by allowing the system to solve it. AI has significant advantages over traditional programming and can help with processing of life insurance and other applications. The basic building block of AI is called the perceptron measured in weights. Multi-layer perceptions are the basis of deep learning and can learn to solve problems.
Think about this…in 2030 underwriting as we know it today ceases to exist for most personal and small business cases. Instead of taking days or weeks to approve, it will take seconds as the process of underwriting is automated and supported by a combination of machines and deep learning models. AI has a place in life insurance: Always “on” customer support via Chatbots could provide 24/7 customer service; fraud detection and prevention; AI assisted underwriting; intelligent claims processing. We are hearing a lot about ChatGPT (generated pre-trained). A large language model (LLM) is an AI model that has learned the structure, semantics and vocabulary of a language, thus allowing it to recognize, summarize and translate and generate text based on knowledge gained during training. ChatGPT has come a long way and allows for natural language processing, with the technology to analyze the information to improve the experience for the customer. It has the capability of data processing, analysis and insights. Language models are pre-trained with one or more techniques, then they are fine-tuned with domain-specific knowledge. These pre-trained language models that are fine-tuned can then be used in multiple ways, including Chatbots – which is a conversational AI. The majority of us have probably used a Chatbot before. Think about when you are on the internet and “someone” asks if they can help you. This is a Chatbot. In insurance, this could provide 24/7 customer support. Another way AI can help is with fraud detection and prevention. It is estimated that insurance fraud costs at least $80 billion annually in the US. By implementing fraud detection and prevention apps, it is predicted that many of these fraudulent claims could be detected.
We have experienced AI-assisted underwriting in many of our companies. Consider your automated underwriting pipelines/models. These are built with AI, including risk score models and predictive models. This “intelligent process” can automate many of the cumbersome steps of the application and underwriting process and cut back on service times. This straight-through process would reduce workflow and allow processing of higher volumes. However, there is still a need for human experience and intuition.
Assisting in claims processing is another way AI can provide faster turnaround times. It has the capability of providing straight-through processing, thus reducing workload, identifying claims that may have a high probability of being denied, and triaging claims based on how much intervention may be needed. AI provides automated and intelligent document processing.
The insurance industry is undergoing a significant transformation due to AI. There are many uses for it and we can fine-tune it for our specific insurance use cases.
Preventing Fraud in the Digital Age
Bill Conners, CEO, ForMotiv
Jena Kennedy, FLMI, ALHC, ACS, CLU, Senior Director of Strategy, LexisNexis Risk Solutions
Mikele Oldani, Enterprise Sales Executive, Choice Screening
Mike Reeves, FLMI, VP, Solutions Consulting, Sureify
We are all aware that there is fraud in life insurance. We have heard about gypsy rings, money laundering schemes, and using a deceased person’s information to apply for insurance. However, there are also areas of vulnerability that are less visible, especially in the digital age.
The key areas of vulnerability in digital transactions include: account creation, account login, transactions/distribution of funds, mobile channels.
Account creation:
• 1 in every 10 account creations DIN (LexisNexis digital identity network) are attacks.
• There is a 77% year over year growth in bot attacks at account creation across these DIN networks as well.
Account login:
• Highest volume of growth across all core use cases in DIN.
• Across the DIN, 1 in 8 password resets are actually attacks.
Payment/funds distribution:
• Highest volume of attacks across all cases.
• There is a 63% year over year increase in volume of mobile attacks across DIN.
Mobile channels:
• Desktop transactions: 25%
• Mobile transactions: 75%
Life insurance transaction channels:
• Desktop transactions: 61%
• Mobile transactions: 39%
The use of self-service centers opens up access to data, and thus increases the chances for fraud. In insurance fraud specifically, global growth of suspected digital fraud from Q2 2021 to Q2 2022 showed an increase of approximately 159%, top insurance-focused fraud type was first party applications; account takeover attempts in 2021 vs. 2020 showed a 34% increase of those reported in companies. Insurance companies are also seeing a rise in “soft” fraud as things move to digital: 1) Instant decision and lab-free applications are becoming critical to winning business; 2) Vulnerability with data that can’t always be validated such as build, tobacco use and health history - especially if waiving exams and labs; 3) Disclosure changes from tele-interview vs. e-interview. Across 10 insurance carriers, it was noted that there was increased tobacco non-disclosure at a blended rate of 0.6%. Build is even worse with a non-disclosure blended rate of >8%. This can be at a huge cost to the insurance companies.
How do we prevent fraud in the digital age? Many applicants are using mobile devices when applying for insurance and want an easy way to apply. Direct to consumer distribution is also a newer concept. How do we know that they are real people? How do we know where these applicants are? As noted above, 75% of fraud from mobile channels is through mobile transactions. One way to do this is tracking IP addresses, as we can tell where they are coming from. We also should be verifying the identity of the proposed insured right at the beginning and not when a problem arises. Perhaps this is creating an account even before they can apply, email verification, twofactor authentication, emailing customer when a change is made.
Closing Ceremony Show, Don’t Tell: What It Means To Be a Motivational Leader
Cassi Chandler, CEO, Chandler Consulting Group LLC
Chandler led an inspirational closing session where she shared her past experiences with the New York Police Department, Assistant Director of the FBI Academy at the Marine Corps Base in Quantico, and Section Chief of Criminal Domestic Terrorism Intelligence. She highlighted how our work environment is ever-changing, and with the new norm, nothing is normal! Our future changes may involve challenges of more pandemics as well as economic, social and political impacts to include global warming. What-ever the change, we as humans need to look at what is changing and explore. This exploring must come from within.
As we explore within, we must know our own power, extend the power and embrace the champion within ourselves. The champion within us allows for individuals to extend our power and build commitment that helps us see, makes us think and moves us to action. We must also inspire hope during bad times, be an anchor for other people, serve others and most importantly BE KIND!
After Chandler's presentation, Jennifer Johnson introduced the incoming AHOU President, Mark Lounds of Munich Re.
Mark your calendars now for the 23rd Annual AHOU Conference May 5-8 in Boston!
NEW FALUS AT THE 2023 AHOU CONFERENCE IN HOLLYWOOD, FLORIDA
Congratulations to all the new FALUs!