Insurance Has Five Problems – And AI Isn’t One of Them

This article is by Prathiba Krishna, (pictured) AI and Ethics Lead for SAS UK & Ireland and one of the authors for  the voluntary AI Code of Conduct for Claims, and Eddie Longworth – Founder & Director JEL Consulting, initiated and facilitated the creation of this code to ensure collaboration and proper usage of AI in claims.

While millions of Brits trust their lives and livelihoods to insurers every day, the insurance sector itself is facing a difficult period.

A snapshot of just some areas of the market shows how tough it’s been in recent years. In 2023, UK motor insurers experienced their worst performance since 2011, with a net combined ratio (NCR) of 112.8%, which is a loss-making figure. And in 2022, UK home insurers also experienced their worst year on record, with an NCR of 122%.

More frequent and severe weather, as well as tough economic conditions, have led to premiums and deductibles skyrocketing. Carriers face significant backlash as policyholders scramble to find coverage in abandoned, high-risk markets. Meanwhile, analysts warn that the rate hikes and non-rate underwriting actions taken to navigate these market conditions will prove unsustainable in the long term.

In search of workarounds, some insurance leaders pinned hopes on the generative AI (GenAI) boom for quick fixes. However, allegations of faulty algorithms rendering unfair denials have fueled critics, who claim that AI has become another problem in an already fraught landscape.

The truth is more nuanced.

While insurance leaders will certainly encounter obstacles as they advance their analytic maturity, AI isn’t the problem. Rather, with the appropriate ethical guardrails and human oversight, trustworthy AI is a solution, delivering the insights and agility needed to redefine the industry.

At this potential turning point, experts from SAS have examined and offered insights on the top five insurance problems. Confronting the insurance community’s current technology challenges won’t only bring AI closer to its full potential; it will also help future-proof the sector. Let’s take a look at those problems in more detail. 

1. Data chaos awaits law and order.

When data points align with private personal details, the current lack of legislation and regulation governing the use of artificial intelligence can feel unsettling – especially for an industry so steeped in compliance and regulatory reporting. Language in the EU AI Act, China’s Interim Measures and the NAIC AI Model Bulletin are among the first efforts to establish AI guardrails in insurance, but with the regulatory landscape in flux, insurers and insurtechs are stepping into the breach with proposals for self-governance.

To set the stage for meeting regulatory standards yet to come, the insurance industry, like the banking sector, must prioritise data lineage and governance within its AI capabilities. As important as it is for insurers to extract valuable insights from their large datasets, it’s equally important to cleanse this data of errors and inconsistencies. This helps ensure reusability, improve decision-making accuracy, boost productivity and reinforce the reliability of results. 

AI education will also be a key determinant in successful AI deployment, and in preparing for future compliance. Fostering data literacy across the organisation empowers the entire enterprise to discuss, understand and ultimately embrace ethical AI practices.

2. AI overload strains risk management.

Amid the industry’s rapid digitalisation and the explosive growth of AI and generative AI, risk managers are rightly concerned about the unintended consequences of robot algorithms – particularly as business leaders race to translate AI productivity gains into long-term business value.

Prototyping may look promising, but production AI requires robust infrastructure to ensure responsible and safe deployment. Meanwhile, “black box” AI solutions that limit customisation may appeal to executives for the perceived simplicity, but the lack of transparency and explainability exposes the organisation to considerable AI risk.

Insurers must delve deeper into the importance of integrating AI into existing systems while aligning with an enterprise AI strategy with strong governance. Insurers must also consider the broader scope of GenAI use cases beyond large language models. Effective applications of synthetic data generation, for example, could strengthen data privacy while optimising pricing, reserving and actuarial modelling.

3. Indemnification tunnel vision stalls progress and partnership.

A paradigm shift is brewing in insurance – that is, if the industry can reach the required critical mass. The tech sector foresees carriers evolving from reactive indemnifiers to proactive partners with their policyholders, consumers and businesses alike.

Consider the following use case: the World Health Organization recently reported a staggering proportion of health events – over 30% of global cancer deaths and 80% of chronic diseases – are tied to preventable habits. Meanwhile, insurers already collect extensive health data on their customers to offer appropriate coverage. Why not put this data to use?

Through existing channels like smartphone apps, insurers could offer customers the opportunity to opt in to AI-powered health coaching, delivering tailored advice that reinvents the conventional customer experience – and reduces policy payouts. Beyond wellness services, insurers should also strongly consider the market potential of partnerships for climate change and ESG. Not only could such initiatives help address insurers’ solvency issues; they could greatly enhance public perception of the industry.

With appropriate ethical guidelines in place, the insurer-as-partner model is far from a fantasy. Cutting-edge insurtechs and parametric insurance policies operate in a similar vein, making this a novel path for forward-thinking insurance players.

4. Hidden digital risks require centralised solutions.

With near-ubiquitous smartphone technology, an insurer today can reach markets wherever wireless exists. For insurers looking to offset premium increases for customers, investing in digital migration helps put modern offerings for modern customers on the market.

However, to compete in this digital marketplace, insurers must offer increasingly individualised products and services to policyholders. The appeal of these tailored offerings and ease of signing up online has lured a slew of potential customers, which has proven a blessing and a curse. 

Carriers are inundated by the volume and speed of applications. Unfortunately, in the race to approve or deny coverage, insurers don’t have enough time to properly research and identify the clientele most likely to commit fraud or who fundamentally pose unwanted risk. Therefore, insurers will generally absorb these customers – and the risk they pose. 

Mounting the technological infrastructure to accurately identify fraud and other threats en masse, still thwarts legacy carriers and insurtechs alike. And both insurance fraud and the underwriting of unwanted risks means an increase in the insurer’s loss ratio and combined ratio, and, ultimately, increases insurance premiums for customers.

A successful digital carrier must orchestrate efforts to garner customers, serve them appropriately and balance the various kinds of risks they impose, ideally in an integrated, cloud-based ecosystem.

When insurers centralise and integrate the tasks of actuaries, underwriters and fraud analysts, it helps ensure the carrier can make a profit with risk-appropriate customers, while serving and protecting customers exactly as they need, and at a price that does optimum justice to all parties.

5. Life insurance: newly vulnerable, but still vital.

Life insurance is a perfect microcosm of much of what plagues the insurance industry today. Life insurers have long depended on commercial real estate for profitability. Since the COVID-19 pandemic, however, the value of these assets has nosedived. This subset of the industry needs to create new opportunities.

When we talk about an insurable interest, not everyone needs to protect an asset like a car or a home – but everyone has a life. Health data organisations forecast global life expectancy will increase to more than 78 years by 2050. With growing risks in a world in polycrisis, life insurers have a part to play in driving positive change.

For example, one of the many forces that can begin and perpetuate generational poverty is a death in the family, and the resulting decrease in income or support. When the worst happens, dependents left behind often face the stark reality of going without. Life insurance can greatly ease the financial burden, but unfortunately, due to a lack of accessibility and historic marginalisation, many who could greatly benefit from a policy are uninsured.

Today, by incorporating cleansed data, with pricing decisions made within principled frameworks, and with the global outreach of digital platforms, insurers can reach, educate, and protect more people, potentially breaking multi-generational cycles of suffering.

Reimagining insurance in context

Addressing the environmental, economic and ethical challenges facing the insurance industry will require human ingenuity. AI and other technologies to power a more equitable and climate-resilient pivot for the sector – and bestow competitive advantages to traditional carriers and insurtechs in the process.

To further explore how insurers can learn, adapt and compete within the current insurance marketplace, download Top 5 insurance problems – and AI isn’t one of them.