Advanced analytics (AA)—including artificial intelligence and machine-learning methods— continues to be a top-of-mind topic among executives in the insurance sector. However, despite significant value at stake—an estimated €1.2 trillion in value globally1—and increased competition from tech-native entrants, the uptake and impact of AA in insurance across Europe, the Middle East, and Africa (EMEA) is varied.
A new study from QuantumBlack, AI by McKinsey has found that while EMEA’s insurers are investing in analytics use cases across the value chain, most fail to capture significant value from AA: 86 percent of surveyed EMEA companies either realize less than 5 percent of their operating profit from AA or are not tracking value capture from AA at all. (For more on the methodology, see sidebar, “About the research.”)
The relatively low value realization is driven by a number of factors, including a relative lack of C-level sponsorship for analytics, insufficient alignment with business for execution, and a relatively slow pace of use case development that is both siloed (thus lacking synergy) and relatively slow (typically taking nine to 12 months per use case). Furthermore, data assets and technical infrastructure are often built without a clear vision or a link to value, and many insurers’ foundations in terms of data, talent, and technology are frail: in our survey, most insurers listed both data and talent among their top five roadblocks to scaling AA. On average, the surveyed insurers aim to nearly double their analytics talent over the coming two years, but most lack a clear hiring, retention, and capability-building strategy.2
Given the current macroeconomic context, the spread in performance among insurers is set to increase, and insurers that accelerate their AA efforts have a significantly greater chance of being among the winners. In this report we further unpack the findings of our research, exploring what sets insurance leaders in AA apart and what other insurers in the region can learn from them.
What the best look like: EMEA’s top performers
Just four companies surveyed qualified as top performers and realize substantial impact from analytics, according to our analysis of company performance across six dimensions indicative of AA maturity (Exhibit 1). The application of AA resulted in a 10 to 25 percent uplift in the four top performers’ operating profit last year, and they expect this impact to increase over the next two years—with some companies expecting to achieve more than a 25 percent uplift in their operating profit as a result of AA.3 Additionally, AA creates value for customers through fairer pricing and more tailored distribution; AA creates value for employees through increased productivity and offloading of repetitive tasks.
Top-performing companies also regularly track and assess the impact of AA and invest more than other companies, with two of the top performers investing between €10 million and €25 million per annum, compared with investments under €3 million for more than half of the remaining EMEA insurers. Most top performers plan to further increase their spending by up to 20 percent per year going forward.
In terms of their operating model, all top performers view AA as a top ten C-level priority, have clear lines of responsibility for driving AA initiatives within the business, and employ cross-functional teams and iterative ways of working. Building on that, top performers are exploring more use cases across the full value chain than other surveyed insurers and are accelerating their use case development by unlocking synergies between use cases within a domain, for example, by using the same teams, data assets, and user interfaces across use cases within a domain, such as motor claims or distribution.
Top performers’ speed of use case execution is also significantly faster than that of average performers in the region, with two reporting a typical use case development timeline of less than three months from exploration to full-scale rollout. As a result, they have an average of six scaled and fully rolled-out AA use cases, compared with an average of two for other insurers that participated in the survey.
Top performers have an average of six scaled and fully rolled-out AA use cases, compared with an average of two for other insurers that participated in the survey.
Top performers have also taken steps to ensure that enablers for AA are in place across the organization. First, they have an attractive employee value proposition and are continually developing and supporting analytics talent. They also employ a larger workforce skilled in AA: they have on average 93 AA in-house and contracted or vendor-deployed AA professionals, compared with an average of 54 among remaining insurers. Second, all have fully adopted machine-learning operations (MLOps) practices and use AA technology, including cloud platforms, to meet a high degree of storage and computing capacity needs.
When it comes to data, top performers are better than most in all the surveyed aspects of data infrastructure and data management—including governance, data quality, data cataloging and discovery, and lineage. However, these and other aspects of data continue to be a top roadblock for the surveyed insurers, including the top performers, and most have clear plans on further improving their governance, data availability, and infrastructure to further scale the value realized from data.
Outside the top: Large spread in analytics maturity across EMEA’s insurers
Outside of the top-performer group, insurers in the EMEA region are investing—to varying degrees—in AA across the value chain, and executives increasingly recognize and seek the value of AA. Nevertheless, only a fraction of the potential impact has been unlocked. At the same time, new tech-powered entrants to the insurance market, such as Lemonade, Root, Tesla, Amazon, IKEA,4 and Verily, are seizing some of the opportunity. This disruption is underlining the competitive advantage of digitalization and creating an increased sense of urgency among incumbents to scale analytics.
As things stand, the maturity levels of most EMEA players are relatively low, especially when compared with the four European top performers. Looking at the composite score across the six dimensions assessed in the survey, used to gauge a company’s overall current data and analytics maturity, the spread between EMEA insurers and within countries is large, with the bottom and the top performers achieving 18 percent and 85 percent scores, respectively (Exhibit 2). Countries’ relative performance also varied, with the average for insurers in the United Kingdom and Spain generally outpacing those in Benelux (Belgium, the Netherlands, and Luxembourg) and South Africa.
None of the 59 participating companies, including the top performers, describe themselves as analytics-driven organizations (Exhibit 3). Most have proofs of concept tested and implemented, with ambition to scale, but only 19 percent use AA extensively for decision making, and three German insurers and a South African insurer went as far as to say they have little to no advanced data or analytics capability in place. In our experience, small to medium-size insurers tend to be especially challenged because their investment levels remain subcritical and their talent bench is usually thin.
Perhaps unsurprisingly, most of the surveyed insurers find the bottom-line impact of their AA initiatives to be up to about half of the impact realized by top performers (adjusting for the companies’ operating profit). And EMEA’s insurers are also less optimistic about achieving their impact aspirations from analytics in the future: half are expecting an AA-induced uplift of just 5 percent of operating profit in the next 24 months, significantly less optimistic than top performers’ expected 10 to 20 percent increase over the same period.
Notably, though, while three out of the four top performers regularly track and validate the impact of their AA initiatives, only a quarter of the remaining surveyed EMEA insurers do so, and 13 percent of EMEA’s insurers do not track the impact of AA at all (others do light or moderate tracking). Measuring impact from AA, while often challenging, is fundamental to its success and fundamental to shifting the focus on scaling. This insight remains true across industries. Based on our data, companies that track impact tend to realize more value from AA.5
Without assessing the potential value from AA and tracking its impact, it may also be more difficult to make a compelling case for significant investments in further AA capabilities and use cases. It will be critical for EMEA insurers breaking the cycle of chronic underinvestment to realize their wider organizations’ ambitions for AA. The €3 million that most participating insurers commit each year won’t be enough.
Four ways to scale the impact from AA faster
It’s clear that AA leaders in the EMEA insurance industry are pulling ahead in impact and performance, and EMEA’s less-mature insurers have an opportunity to emulate their success. Our survey and our client experience highlight four themes common to companies that successfully apply AA at scale: first, they follow a value-backed strategy; second, they apply a domain-centric approach in expanding their set of AA use cases; third, they align with the business for use case execution; and fourth, they systematically invest in the enablers that matter: data aligned with use cases, talent strategy, and fit-for-purpose tech and tooling.
Follow a value-backed strategy
Assessing the value expected from AA initiatives is essential not only to inform the prioritization of AA use cases but also to justify the investment into AA use cases, data, talent, and tech. Many insurers today invest in extensive data technology and cleanup exercises without a specific list or sequence of the use cases to pursue, or, despite having strategies and use case road maps in place, they often can’t map these back to value easily.
By contrast, the most successful players subscribe to a top-down, value-backed approach to data and analytics: they link analytics initiatives to clear business strategy objectives, prioritize them by measurable value, and build out the required data assets and technical infrastructure in line with those priorities. The use cases are also driven by business demand and clearly linked to overall financial targets—that is, “baked into the budgets.”
A successful analytics business strategy centered around maximizing value entails starting with use cases expected to generate the highest value and business impact and building reusable data foundations “as you go,” with data pipelines and assets built for reuse in other use cases. Avoiding gold-plating data standards and data management and implementing a strong collaboration among business, data, and IT—with the business responsible for driving the implementation—is also critical. Using the principles outlined above to focus on the few use case areas (potentially within a single domain at a time, as outlined in the next section) that are likely to move the needle on value, as opposed to launching dozens of initiatives at the same time, is a highly impactful approach. Once the highest-value use cases have been implemented, insurers can move on to explore nascent use cases that are more exciting, which top performers are currently doing (such as usage-based products that use Internet of Things data, preventive usage of sensor and habit data for illness detection for health and life insurance, and highly customizable products, such as microcoverage).
Scale analytics use cases in a domain-centric fashion
When it comes to use case deployment, more is better. Companies that deploy more use cases tend to have a higher overall analytics maturity (r=0.45)6 and better performance. The key question is how to increase use case deployment quickly and efficiently.
To date, the application of AA in insurance has typically involved the development of individual high-impact use cases in discrete domains. Our survey results reflect this: on average, EMEA’s insurers tend to focus on one or two use cases per domain, with the most scaled examples in cross-selling and upselling in distribution and fraud detection in claims (Exhibit 4). If successful, these blockbuster use cases are then scaled across the wider organization and their insights are embedded into the front line. With the hype of a few successful deployments, use case development often proliferates, typically with siloed teams working on different use cases in parallel. This is inefficient, increases complexity, and often doesn’t align with the value-backed approach.
To fully capture the potential of AA, insurers may need to move beyond the blockbuster approach and take a holistic, domain-focused perspective, tackling a portfolio of use cases in each domain. AA leaders have demonstrated that there are significant synergies in such a domain-centric approach that can accelerate use case development and unlock greater value for insurers. For example, data used for claims triage and severity prediction can be reused in repair assessment and claims assessor prioritization, while also providing valuable data to management and the business to drive value in claims. A portfolio approach can also significantly transform the functions themselves. In pricing, for example, this can involve massively enhancing the capabilities of the actuarial team, streamlining governance structures, and raising the appetite for—and the metabolic rate of—ingesting new data sources.
Align with the business for execution
Great model outputs don’t provide business value unless the execution is right. Alignment with the business, both on the prioritization of use cases and on their execution, is therefore essential. For example, cross-sell or next-best-action models require business involvement to design campaigns and train frontline staff to act on leads, and senior-business sponsorship is key in facilitating the required behavioral change. Moving quickly from piloting to deployment is critical for many of EMEA’s insurers—which, despite having fewer fully rolled-out cases than the top performers, pilot approximately twice as many—and this requires not only productionizing models but also, crucially, business involvement in mobilizing the front line and users.
Broadly, a successful execution model contains two key elements. First, executive-level accountability and senior-business sponsorship of AA are needed to remove obstacles within the business, increase transparency and agility, and ensure the relevant resources are in place. In our study, all top performers reported that they have a clearly defined leader for AA, but this is lacking in about a third of the remaining insurers (Exhibit 5). At two of the top-performing insurers, the AA champion is a C-level executive, whereas only about a quarter of remaining insurers have AA leadership at C-level.
The second vital business shift required to support AA transformation is setting up combined business and analytics teams and implementing iterative ways of working (Exhibit 5). Additionally, employing an iterative approach to work with a “test and learn” mindset can significantly boost speed of delivery. Roughly one-fifth of the surveyed insurers consistently employ an iterative approach to work, compared with 100 percent of the top performers.
Roughly one-fifth of the surveyed insurers consistently employ an iterative approach to work, compared with 100 percent of the top performers.
Cross-functional use case teams need domain experts fully dedicated to providing business context, aiding in the interpretation of data and the results, and, most important, defining the best piloting and implementation approach. Ideally, domain experts work hand in hand with data and analytics colleagues from day one of use case development, and their role can include designing campaigns, enforcing the required change management, managing communications, owning the relevant sales processes, and training the frontline users of AA solutions.
One role that is particularly critical to making cross-functional teams effective is that of business or analytics translators. Akin to digital-product owners, skilled analytics translators translate business needs into modeling output, ensure that business inputs are made into the analyses throughout (for example, by guiding exploratory data analyses and participating in discussions on defining target variables), and interpret model outputs to business stakeholders. Our study indicates that three out of the four top performers consider analytics translators to be key to the success of analytics initiatives.
Systematically invest in the enablers that matter
To achieve their AA ambitions, EMEA’s insurers will also need to invest in the enablers in which they lag behind top performers today. This includes building on their foundations in data and investing in truly reusable data assets within and across domains, making sure data is fit for purpose in AA; investing heavily in recruiting and AA talent retention and capability building; and continuing to revamp their tech platform, shifting to cloud, and embracing open-source tooling.
Rethinking the approach to data. Ensuring data reliability, data quality, and modern data architecture is the top self-reported roadblock to further scaling AA across the companies we surveyed, including the top performers. Three of the four top-performing companies, and 56 percent of the remaining insurers, listed data among their top five roadblocks.
In terms of data governance—including the management of data quality, cataloging and discovery, and data lineage—most companies, including the top performers, have room for improvement (Exhibit 6). Despite top performers achieving a score of 3.8 on data governance—compared with the survey average of 3.3—most companies, including top performers, describe themselves as only moderately mature in terms of data governance. These companies are exploring basic cataloging techniques and have dedicated teams working on data management and quality but lack well-defined processes and sufficient automation. Most participating companies also have clear plans for improving their data governance. External data partnerships can also be enhanced: most companies use external data on an opportunistic basis but have no systematic process in place and plan to expand external data providers only on a case-by-case basis.
To move to the next level with respect to data, it is essential to start with a data vision including data prioritization according to the insurer’s value-generating potential and business needs, clearly specified goals, and defined accountability and ownership. Organizations also need to ensure good management of the relevant metadata and data lineages, put in place processes for data quality control at the source, and set policies, standards, and processes to describe governance. Adhering to data privacy and security guidelines is of course nonnegotiable.
Finding the right talent. A shortage of data and analytics talent is seen by one-third of the surveyed insurers as one of the top challenges in developing and scaling an effective AA strategy, and many EMEA insurers perform moderately or weakly on various measures of AA talent (Exhibit 7). Average EMEA players also currently have substantially smaller analytics workforces than top performers, comparatively: an average of 54 versus 93 AA professionals per company (including both in-house and contracted or vendor-deployed professionals). Most players are aiming to significantly grow their workforces: the average surveyed insurer is planning to grow by roughly 70 percent in the coming two years—much faster than top performers’ growth plans of roughly 30 percent.
To succeed in this, EMEA’s insurers can advance their analytics recruitment approach by enhancing their employer value proposition—highlighting interesting problems to work on, opportunities for career progression, and apprenticeship opportunities, while offering modern tech stacks and tooling—and reviewing their recruiting channel strategy, going beyond traditional channels such as LinkedIn to explore alternatives, including technical online platforms such as Stack Overflow, and academic conferences, such as NeurIPS and ICML. Recruiters can also access pockets of talent by developing partnerships with, for example, diversity groups, such as Women in Machine Learning (WiML) and LatinX in AI (LXAI), and by engaging in indirect efforts, such as by publishing technical content on content platforms or through podcasts. Insurers can also support their recruitment with dedicated resources, such as experienced and fully dedicated digital and analytics recruiters, and can set up programs or academies focused on capability building for technical talent to help excite and retain talent.
Insurers can also focus on increasing talent productivity to decrease the amount of resources required and reduce the pressure on talent sourcing. This can be done by exploring new ways of working and sharply increasing levels of automation. For example, prepackaged analytics pipelines designed to deliver insurance AA use cases, such as InsureX, require smaller than usual AA teams while also accelerating and derisking use case delivery.
Investing in technology and tooling. Implementing a successful AA strategy requires a strong technological backbone; all top performers included in the study have advanced technological foundations such as cloud infrastructure and MLOps tooling, develop and leverage reusable AA products, and continually keep pace with the latest machine-learning techniques. Despite a moderate score on technological maturity on average (Exhibit 8), 16 percent of the remaining EMEA insurers self-identify as low in technological maturity, noting outdated technology and limited flexibility.
Additionally, top performers show higher use of machine-learning software and seem to be adopting advanced tooling faster than other insurers (Exhibit 9).
EMEA’s insurers have, by and large, laid the groundwork for AA within their organizations, but most have not yet tapped its full potential. Most are piloting use cases, but unfortunately this is where many are stuck. To get beyond this pilot purgatory and capture a clear competitive advantage from AA, insurers in the EMEA region will have to commit to scaling up a larger number of use cases across the value chain on a domain-by-domain basis. This is a complex task, given that the entire organization must adapt—from its operating model to its talent bench—if integrating AA is going to be successful.
To navigate this highly complex and fast-changing area, companies need to continuously challenge their AA strategy and thinking, looking externally to the latest trends and benchmarks (such as those in this study) to ensure they are truly unlocking the potential of AA and the capabilities at their disposal. For insurance companies at a more mature stage of development, this can take the form of a review of the AA road map with key business leaders in each domain, honestly and comprehensively evaluating AA’s impact to date, and determining what more can be done to push pilots through to production and generate more value from embedding AA more broadly across the front line. In parallel, a review of the AA capabilities and enablers, and in particular considering how models are refreshed and maintained, is usually critical.
In today’s economic climate, future growth is not guaranteed. The next few years are likely to be crucial in determining if insurers can capitalize on their initial investments in AA and scale sustainably. Those with a clear path to scale will be more likely to take their place at the head of the pack and help shape the future of insurance in the region.