How Data & AI are reshaping the future of insurance

Introduction

According to Deloitte’s State of AI in the Enterprise research, 94% of business leaders agree that AI is a critical success factor in the next five years [1]. Although this view is also embraced by Insurance leaders, most firms in the industry are still in the early stages of their Data & AI maturity [1,2]. To enable insurers and other organisations to reach higher maturity stages, Deloitte has developed the Insight Driven Organization (IDO) framework. Becoming an IDO is about augmenting, accelerating, and enhancing decision making with technology [3]. It is about creating more efficient and effective operational processes, and about improving customer relationships through seamless and personalised experiences. To become an IDO and unlock the full potential of Data & AI, insurers have to adopt a holistic view across five dimensions: strategy, people, process, data and technology.

In order to reach a higher Data & AI maturity stage it is essential that insurers choose the right use cases [1]. Focusing on use cases that are too much of a challenge, or do not generate sufficient benefits in the short-term, might slow down the transition towards an Insight Driven Insurer. In a series of articles we will discuss four use cases that bring value to the organisation. This first article is about pricing.

Pricing: the most important lever for profitability for an insurer

Pricing is one of the most important profit levers for businesses. It can significantly boost an insurer’s profitability [4]. However, it is a complex process that requires a holistic approach across the five dimensions of the IDO framework mentioned above. In a typical engagement we start by interviewing key senior stakeholders and pricing experts to assess the current state on each of these five dimensions. As a next step, we facilitate a workshop with senior leadership. The purpose is twofold. First, we share and discuss our findings to validate and to finalise the current state assessment. Second, we present an outside-in perspective based on our (inter)national experiences and extensive knowledge of the Insurance market. This outside-in perspective is the start of a discussion to define the ambition and vision on pricing and to align it with the overall business strategy. Finally, we make a gap analysis regarding the current and target states and we build a roadmap to determine what is required on all five dimensions mentioned before to bridge the gap. In the remainder of the article, we discuss typical challenges that insurers face and suggest how to solve them.

Getting the price right starts with a clear strategy

In our experience, insurers often do not have a (clear) pricing strategy. They do not have a clear view on how to position themselves in the market, how to differentiate themselves from their competitors, and what portfolios, products and segments to focus on. As a consequence pricing does not necessarily contribute to an insurer’s strategic objectives. However, clear answers to these questions supported by KPI targets are essential and provide necessary guidance to pricing teams, allowing them to align the pricing strategy to the insurer’s overall goals. Answering these questions is therefore generally the starting point of our conversations with senior leadership on defining the ambition and vision on pricing.

Decision making & governance need re-thinking

Each pricing adjustment at an insurance company requires the approval of many stakeholders who do not have clear KPI targets. This extensive decision-making process substantially slows down the time-to-market of new pricing models. To improve agility in responding to changes in consumer behaviour, market conditions and the competitive environment, we advise insurers to adopt a leaner and meaner decision-making process. Having a pricing governance in which decisions are supported by a transparent framework and assessment criteria, enables pricing teams to have a clear and strong mandate to adjust prices. Only pricing adjustments that have a major (expected) impact on an insurer’s bottom line would require the approval of all stakeholders.

Pricing model innovation & investment in Data&MLOps

Insurers traditionally rely on historical data from claims and policy management databases, enriched with aggregated customer data, to predict future claims and set commercial prices. Partially due to a complex IT landscape, actuaries and pricing analysts spend a significant amount of time on data processing to obtain a dataset that works for them. On top of that, insurers generally approach pricing – including data processing and modelling – as one-time projects. This approach creates unnecessary inefficiencies and re-work in the pricing cycle, leading to a longer time-to-market. Furthermore, there are still insurers who do not use conversion, retention and customer lifetime value models for their commercial pricing. In the latter case, we advise insurers to start developing such models to catch up with the leading practices in the market. As a next step, once such models are in place, insurers must learn to continuously capture value from data at scale. This can be achieved by applying DataOps and MLOps, which are approaches to industrialise AI. Both DataOps and MLOps are combinations of technological solutions and organisational methods to streamline the development and delivery of data, data science and machine learning projects in a reliable and efficient way, using an iterative process of short delivery cycles. Preferably, most elements in these cycles are automated. To achieve excellence in pricing today, insurers must start to invest in the optimisation of their pricing processes.

To achieve pricing excellence in the future, insurers need to broaden their horizon beyond claims and policy data. Developments in, for example, Internet of Things, XaaS, ecosystems & platforms, and autonomous vehicles will have a significant impact on the insurance industry and will generate a totally different type and amount of data. Insurers must start to collect and work with these new types of data, which will be provide more insight into the behaviour of their (potential) clients, which can be used for demand modelling, risk assessment and pricing. It will also provide early adapters with valuable experiences working with non-traditional data, and as a consequence non-traditional modelling approaches based on AI and machine learning. In the longer term this will offer them a strategic advantage when the trends and AI driven transformations become more prevalent.

Technology that allows you to develop & deploy increasingly sophisticated pricing models

Once a new pricing model has been approved, it currently takes a substantial amount of time before it is deployed to the production environment. Generally, the pricing model needs to be exported from specialised pricing software into an Excel file and then manually entered into another IT application that is part of the production environment. The risk of mistakes is high, and extensive testing procedures are in place to ensure the right price is entered into the production environment. Insurers can and must change this practice, and some frontrunners have already done so. There are several solutions available in the market that enable pricing models to be easily deployed into the production environment – having in place necessary checks and balances to avoid costly and embarrassing mistakes. As a consequence, these insurers area able to adapt more quickly to shifting consumer demand, market conditions and competitive environment. As AI allows for more advanced pricing models and real-time optimisation, and current practices become essentially infeasible, a smooth deployment process will be even more crucial.

A new people approach to broaden the skills of more traditional pricing teams
Optimising pricing by means of Artificial Intelligence is as much about data and technology as it is about people. To improve their chances of success of an AI-transformation in pricing, insurers must enhance their pricing teams with AI talent: data engineers, machine learning engineers, and data scientists. Adding these competences offers insurers an opportunity to establish new ways of working which are more common in AI-based organisations: multi-disciplinary teams that embrace a culture of shared responsibility and collaboration and have end-to-end ownership of their pricing models.

The time to act is now

As pricing is the most important profit lever, insurers have a natural incentive to further improve their capabilities in this area. Apart from the holistic approach described above, there are several ways in which we can support insurers on this journey. For example, based on our knowledge of different pricing solutions and their providers we can support the selection of the technology that fits your organisation best. Alternatively, we can assess the governance and (approval) processes and suggest avenues to make them leaner and meaner while maintaining adequate risk management. Similarly we can assess the data and models currently used and recommend areas of improvement. The most successful insurers empower their pricing capabilities with a holistic view across the five dimensions mentioned before: strategy, people, process, data and technology. By investing in their pricing capabilities, insurers can quickly respond to changes in consumer behaviour, market conditions and the competitive environment. As these are times of high inflation, economic uncertainty and possible regulatory changes, it is time to act now.