Generative AI (GenAI) has profound implications for all industries. According to PwC’s Global CEO Survey 2024, 70% of CEOs believe that GenAI will significantly change the way their company creates, delivers and captures value. Among the insurance CEOs surveyed, 28% expect that GenAI will decrease their headcount by 5% or more in 2024, the second highest level of expected change among all sectors.
Industries most exposed to GenAI (Top 10)
Behind these headline figures lies a struggle to identify the optimal uses for GenAI – to cut through the buzz and establish where and how to invest. With an array of choices in front of them, executives confront the critical challenge of not just what to choose, but when to implement it and who to partner with – decisions that carry significant strategic weight.
We believe the following four key steps can help insurers navigate the GenAI revolution, and we offer examples of how insurers are already deploying GenAI with success today.
Step one: Set the AI strategy
One of the most common pitfalls that insurers run into when first starting out with GenAI exploration and experimentation is pursuing use cases without a clear strategic direction. Setting an overarching strategy for AI and automation beyond GenAI alone ensures a cohesive technological ecosystem that maximizes synergies between different AI technologies (such as RPA, predictive analytics, image recognition and conversational AI), preventing missed opportunities and strategic mismatches. This broader perspective fosters an integrated approach to use case selection and prioritization, ensuring all technological investments are aligned with the company’s long-term objectives and can adapt to evolving market demands and technological advancements.
Example
We supported a midsized insurance player in the US to embark on their GenAI journey by establishing clear guiding principles that any future AI investment would need to be evaluated against to simplify decision-making. These included: Maximizing automation of non-core processes, leveraging reusability of patterns at all times, committing to and ensuring ethical standards across all AI deployment, optimizing for system integration and interoperability with no legacy build-ups, and focusing on flexibility and scalability over ROI.
In our experience, the firms that succeed and excel with GenAI follow a well-defined strategic path informed by a clear vision of their future state and understanding of their organizational readiness.
Step two: Double down on GenAI in prioritized capability areas
Once an insurer’s (Gen)AI strategy is set, they need to make a distinction between the types of capabilities in the organization where GenAI could be applied. That starts with the foundational elements that the business needs to operate (‘lights-on’ capabilities) where the focus should be to ruthlessly aim for cost levels below competition, and those that are needed to compete in the market (‘table-stakes’ capabilities), where the focus should be to cut costs while keeping the balance to maintain ‘good enough’ quality.
These two sets of capabilities are attractive targets for GenAI implementations, to help redirect resources and strengthen the strategic focus on capabilities that build a competitive advantage (‘differentiating capabilities’). Typically we find that insurers have a cost base that is excessively skewed towards table stakes, lights-on or even capabilities that are not required at all, as opposed to the differentiating capabilities that drive meaningful results and value. GenAI can help cut and redistribute those costs.
Optimal vs. typical investment allocation
Investments in GenAI should be used as a lever to free up and redistribute critical resources toward differentiating capabilities that drive meaningful value for clients and strengthen the company’s competitive position.
Example
A leading Dutch insurer is experimenting with a GenAI chat capability to radically simplify policy language for clients and colleagues. Such GenAI-powered chat capabilities are relatively easy wins because they tend to be industry-agnostic and easy to implement – deploying them in non-differentiating areas of the business has allowed insurers to get familiar with GenAI deployment without taking disproportionate risks.
A ruthless capability assessment can identify non-differentiating capabilities where insurers can deploy GenAI to free up resources and reallocate to parts of the business that drive meaningful client and competitive value.
Across the majority of insurers, we find that 50% to 70% of all activities that are characterized by high volume and low variability can be labeled as non-differentiating table stakes and lights-on capabilities. These are promising potential targets for GenAI implementations – indeed, our research has already shown that GenAI will be most impactful in non-differentiating areas such as policy generation, claims, and customer service.
A GenAI use case / Impact breakdown
Step three: Craft the GenAI blueprint
A GenAI blueprint identifies the specific table stakes and lights-on activities within chosen focus areas that can drive disproportional upside. While GenAI has broad applications, deployment of GenAI in operational processes typically is most effective when used for specific activities with clear goals. For that reason we recommend choosing a few main areas in the value chain where insurers want to focus their GenAI deployment.
Examples
A smaller P&L insurer deployed a personalized AI policy generation tool that considers customer input and existing customer data, including factors such as demographics, risk profiles and historical data, to generate customized policies with dynamic and personalized premiums tailored to each individual customer, which led to around a 10% operating profit uplift.
A midsized P&L insurer deployed a GenAI-powered conversational broker chatbot, recommending insurance products to consumers and smaller corporate clients and providing personalized insurance recommendations, which resulted in an operating profit uplift of 5% to 10%.
Underwriters at a midsized European insurer now use a conversational AI chatbot to query portfolio data in order to ascertain risk appetite or make decisions on new product launches. This insurer successfully realized around a 5% operating profit uplift.
In our experience, insurers are most likely to deploy GenAI in areas that demand a high volume of both structured data (census files, claims history, and client relationship history) and unstructured documentation (such as claims notes, call transcripts, and broker emails). The good news is that these are prime candidates for GenAI to demonstrate its potential, by extracting key pieces of data, summarizing information succinctly and improving efficiency.
However, use cases are not a strategy in themselves – they must be chosen and prioritized to match the business strategy and future vision for GenAI identified in steps one and two. Working with scenarios allows insurers to simulate and assess the combined impact of multiple GenAI use cases, enabling them to strategically prioritize those that enhance operational efficiency, customer experience, and competitive advantage. This holistic approach ensures resources are allocated to initiatives that promise the most significant overall impact.
Additionally, to ensure alignment it may be necessary to engage external parties to understand the applicability of a potential use case to the insurer’s own operations, and project the extent to which the GenAI solution can reduce, simplify, or eliminate specific workloads within priority areas.
Step four: Partner with the right GenAI allies
While the explosion of interest in GenAI has been closely associated with just a few public solutions such as ChatGPT, public solutions are actually the least commonly used across industries. Within the insurance sector, the vast majority of use cases (around 60%) will be implemented using third party off-the-shelf solutions. Where new products, services and personalized experiences are concerned, around 20% of implemented GenAI solutions will be bespoke solutions. This means that partnerships will be critical to the success or failure of many insurance industry GenAI strategies.
Among current GenAI providers of size (those that have raised in excess of $20 million), applications in insurance are concentrated around sales and marketing, operations (primarily claims) and customer service. However, since most providers are industry-agnostic it is especially important to validate the potential partners’ understanding of the industry and specific situation within which their solution will be deployed. Operational applicability, value captured, partner benefits, and colleague or client impact will be the key measures.
Our research suggests that the most successful partnerships are those where both the insurer and the GenAI provider have clearly delineated their roles and responsibilities across key functional building blocks and multiple use case applications, with a regular periodic review of impact to date, progress and lessons learned.
Conclusion
It is worth restating the essential elements of a GenAI strategy that is focused on real world value: The first element is to define a strategic AI vision, one that differentiates the meaningful from the merely novel. Then, focus on key capabilities – those that differentiate an insurer’s business, sustain or improve competitiveness, or support foundational operations. Thirdly, develop a detailed GenAI blueprint, identifying the most impactful use cases and the capabilities needed to implement them. Finally, choose the right partners – companies that can align with the company’s long-term vision and evolve with its needs. At each step of the AI journey, insurers should consider how the business can leverage best practices – such as setting an overarching strategy for AI and automation beyond just GenAI, and using scenarios for use case selection and prioritization – to improve the chances of success.
These are the building blocks of a GenAI strategy that delivers value. It is likely that any insurer will find that partnership in the widest sense is the most critical element of all: The GenAI age is uncharted territory, and in uncharted territory a guide makes all the difference.
Contact us
