In today’s rapidly evolving insurance landscape, the arrival of AI in Life and Health underwriting and claims workflows is no longer a distant possibility – it’s already a reality.
Insurers are now investing in AI with the expectation that it can transform underwriting and claims, improving both efficiency and customer experience. Indeed, a boost from AI could not come at a better time – Swiss Re’s 2024 UK Life and Health Underwriting and Claims Watch revealed that underwriters and claims assessors are reviewing ever-growing volumes of cases, often with a sizeable amount of evidence to review for each of these cases. AI can help automate repetitive administrative and labour-intensive tasks, such as summarising documents, enabling more efficient case reviews and allowing additional time to address complex, high-value cases.
While AI can improve underwriting and claims processes, it may also produce biases and errors. Therefore, human oversight and transparent communication with users about AI use and its limitations are essential. Additionally, insurance executives considering how AI can support underwriting and claims must ensure compliance with relevant regulations and uphold ethical principles such as transparency, fairness, and accountability.
Given that investing in AI technology requires considerable time and resources, insurers must weigh up in which areas, and at what stage, AI will deliver the most value for their specific business needs.
As Life and Health insurers consider potential use cases for investing in AI, here are 10 areas to consider:
- AI for summarising claims and underwriting evidence
By incorporating AI into underwriting and claims processes, insurers can reduce manual effort and improve triage, thereby streamlining workflows and facilitating collaboration between departments. Underwriters and claims assessors spend a significant amount of time summarising lengthy medical reports. In the UK, compiling a comprehensive GP report may take a senior underwriter or assessor 45 minutes to an hour or more to summarise. Generative AI can reduce this time by condensing and structuring extensive reports into brief summaries that highlight key risk areas, enabling underwriters and assessors to concentrate on decision-making, which may decrease overall processing times while allowing underwriters and claims professionals to simply validate these summaries, saving time.
- AI to ease access to underwriting manuals
Underwriting and claims manuals often contain detailed information concerning various disease impairments, products, and ratings, and typically require technical expertise to interpret. Extracting relevant information from manuals, including rating calculators, can be time-consuming. Life Guide Scout is one such example, where generative AI enables underwriters to access the content in the Life Guide underwriting manual in a conversational form.1
- AI for supporting customer service
Generative AI, which has already seen extensive use in customer service within sectors like the airline industry, can be leveraged through advanced chatbots trained to handle a wide range of complex queries. This allows only the most nuanced or exceptional issues to require human intervention. The same functionality can be used to support insurance applicants with some technical queries which could easily be handled by chatbots, leaving more complex enquiries to be handled by human customer service agents.
- AI’s role in navigating insurance fraud
Insurance fraud remains a persistent challenge across the industry. By rapidly analysing large datasets, AI and machine learning can detect patterns indicative of fraud, aligning suspicious claims with established risk indicators. In addition, predictive models assist in the early identification of potential high-risk claims, enabling proactive risk mitigation.
For instance, AI can highlight underwriting or claims scenarios resembling previous fraudulent or high-payout cases. This gives claims teams the chance to decide what steps might be needed, such as engaging with the customer or sending the case for further investigation.
- AI as a partner to better target medical evidencing requirements
Traditional medical underwriting grids often rely on age and sum insured, adopting a uniform approach. As AI and machine learning evolve, these technologies can learn from historical non-medical limit (NML) outcomes to better predict which applicants may require medical evidence. This predictive functionality reduces dependence on broad, one-size-fits-all procedures and streamlines the underwriting process.
By identifying applicants with a higher pre-test probability, AI enables underwriters to focus their efforts where the collection of evidence will most significantly make a difference. Conversely, for applicants deemed lower risk, the need for supplementary evidence may be eliminated, expediting processing times and enhancing the overall customer experience.
Other areas could be predictive of specific risks, such as propensity to smoke, which could make cotinine tests more targeted and reduce unnecessary medical evidence spending.
- AI to better target post-issue sampling
Historically, post-issue sampling within the industry has relied on random selection. The insurance sector, however, is experiencing a shift towards targeted approaches, utilising advanced models to enhance the detection of misrepresentation. AI and machine learning further refine this process by accurately pinpointing cases at higher risk of non-disclosure, thereby reducing costs and increasing detection accuracy. These AI-powered models support both claims and underwriting risk assessment by leveraging historical data to detect potential misrepresentation, efficiently triage straightforward cases, and flag those requiring more thorough human review. This facilitates quicker resolutions and ensures prompt payment of valid claims.
- AI as an auditing assistant
Underwriting and claims departments dedicate substantial resources to auditing cases to maintain quality and robust risk management. This involves detailed reviews of decisions to identify inconsistencies or deviations from established protocols.
The integration of real-time machine learning can streamline this task by promptly recognising discrepancies as they arise, thereby bolstering efficiency and consistency. Machine learning algorithms enable rapid analysis of extensive data, highlighting issues that might otherwise escape human auditors.
This means that quality checks could be done in real time, enabling quicker corrective action. This allows auditors to concentrate on strategic initiatives, including the development of enhanced risk management frameworks, optimisation of claims/underwriting processes, and adherence to regulatory requirements.
- AI as a sparring partner for evolving underwriting and claims philosophies
Underwriters and claims professionals are frequently required to synthesise information from disparate systems and data sources to fully understand the impact of evolving underwriting and claims philosophies on portfolio performance.
The emergence of generative AI, in conjunction with the implementation of unified data lakes, is now transforming this landscape by enabling real-time learning and seamless integration across platforms.
For example, information derived from claims can be instantly analysed and fed back into underwriting models, allowing for timely and informed adjustments to philosophy and practice. By continuously updating risk assessments and pricing strategies in light of the latest data, insurers are empowered to respond proactively to emerging trends and anomalies.
- AI as a tool to facilitate risk assessment
By analysing vast amounts of individual and contextual information from various sources, such as benchmarking against historic records, pulling in electronic health records, wearables, and historical claims information, AI systems can detect patterns and factors that influence each person’s level of risk. This means AI can tailor assessments to reflect unique circumstances, enabling more personalised risk assessments. This approach enhances accuracy and minimises errors, while delivering a streamlined, efficient onboarding experience. Nevertheless, it remains essential to consider the relevant legislative framework to ensure compliance with applicable AI regulations.
The resultant efficiencies from AI-enhanced processes can enable teams to concentrate greater efforts on strategic, high-value cases – such as designing robust risk management frameworks, optimising process flows, and ensuring alignment with regulatory requirements. Ultimately, these developments enhance the insurer’s ability to deliver fair outcomes to customers, strengthen competitive advantage, and ensure the organisation remains at the forefront of innovation in a rapidly evolving market.
- AI as a tool for continuous improvement
Underwriting and claims departments can also leverage AI to nurture a culture of continuous improvement. In one example, an insurer may implement an AI-powered auditing tool that reviews underwriting or claims decisions in real time. The system flags inconsistencies or deviations from best practice, instantly notifying the underwriting and claims team. Each flagged case is then discussed collaboratively, allowing the team to learn from mistakes or identify where processes could be streamlined. Over time, the department builds a repository of lessons learned, which informs training sessions and updates to underwriting guidelines. This continuous feedback loop not only improves accuracy and efficiency but also encourages a proactive approach to risk management and professional development of staff.
In summary, the next wave of AI is here and promises greater efficiencies for underwriting and claims professionals. The integration of AI into these functions can enable real-time data analysis, personalised risk assessments, and streamlined processes, empowering insurers to enhance quality control, respond proactively to trends, and focus on strategic initiatives. Well-designed AI systems can enable professionals to focus on what matters most – delivering smarter decisions and better outcomes for customers.
Amidst these technological advancements, it is also crucial to maintain a human-centric perspective, carefully weaving the human-in-the-loop into human-AI workflows to ensure expert oversight and ethical decision-making. Human judgement remains vital in interpreting complex scenarios, validating AI outputs, and addressing nuanced cases that require empathy and contextual understanding. Moreover, strict adherence to applicable regulations and responsible AI principles is essential to uphold transparency, fairness, and accountability, safeguarding both customers and the organisation from unintended risks.
Insurers must also carefully assess where and when to invest in AI to maximise return on investment. Those who engage thoughtfully with AI’s potential – balancing innovation and responsibility – will likely lead the charge in a rapidly evolving industry.
Febby Mulewa is Head UW & Claims Portfolio, Market Units L&H Reinsurance at Swiss Re
Maura Feddersen is Behavioural Research Manager, CUO L&H Reinsurance at Swiss Re
You can find the original version of this article, which is reproduced here in full, at https://www.swissre.com/reinsurance/insights/10-areas-to-watch-for-ai-in-claims-and-underwriting.html.
Footnotes
Any views expressed in this article are those of the author(s) and may not necessarily represent those of Longevity & Mortality Investor or its publisher, the European Life Settlement Association
