Press release

Zurich |

The top AI trends for the insurance industry in 2021

Everyone is talking about artificial intelligence (AI). In the business world, the scientific community and society alike, AI is the subject of widespread and heated debate. For companies in the insurance industry, it can be hard to stay on top of the latest developments. While technological innovations are opening up new possibilities at every turn, the use of AI in day-to-day business often lags behind. Yet AI offers valuable opportunities and increases the ability of companies to compete.

What does 2021 hold in store? There is no one-size-fits-all answer to the question of how individual companies should use AI. Each sector faces its own challenges, conditions and goals – especially against the backdrop of the coronavirus pandemic and its impact. adesso sees the following trends for the insurance industry in the year ahead.

Trend 1: explainable AI (XAI)
Trust is an insurance company’s most valuable asset. Upholding that trust necessitates the responsible use of new technology, especially AI. With AI, algorithms rather than humans make decisions. The use of XAI allows people to comprehend machine-based decisions. In many application scenarios, such as claims management and fraud detection, employing XAI is a prerequisite for the use of AI, as such methods are subject to accountability standards. The challenge lies both in technical implementation and making it possible for all involved to understand what is going on.

Trend 2: personalisation
Big data makes policyholders increasingly transparent, with the adaptation of certain processes along the value chain leading to individualised customer service and support. Analysing use and user data allows insurance companies to develop a target-group model for optimised customer communication in marketing and sales. It also opens the door to customised behaviour-oriented adjustment of insurance premiums and benefits, as well as specific forecasts for an approach to providing advice and coverage that takes all aspects into account, to name just a few examples.

Trend 3: approaches
Even before digitalisation projects get under way, customers want to achieve a clear understanding of the process-related and/or monetary added value that can be expected once a technological solution has been introduced. Ideally, that added value should be expressed as a return on investment (ROI). As a result, the importance of a needs-oriented (i.e. specific, application-related) approach to rolling out AI projects at insurance companies is on the rise. Generally, the willingness of insurers to decide to introduce AI solutions appears to rise sharply once it becomes possible to quickly demonstrate initial results and successes in individual areas. This means that insurers want to digitalise selected departments, processes or products with the help of AI first – giving them a proof of concept – before subsequently digitalising entire segments. The modularity required to this end needs to be looked at from a variety of perspectives, such as the professional, technical and price-related modularity.

Trend 4: quality promises
Insurers that roll out an AI solution want to be able to rely on process stability and the quality of the data provided. The collective understanding of what AI can do and where there are physical limitations is increasing. As a result, the expectation that an AI solution will be able perform miracles without any input or guidance whatsoever has been replaced by the realisation that all AI needs to be pointed in the right direction, especially at first, in order to automatically achieve the desired outcomes in future. Generally, insurers appear to be particularly keen on accepting and actively demanding compliance with the quality promises contractually pledged by the operators of AI solutions. A targeted approach and a quality promise increase insurers’ acceptance of AI projects significantly.

Trend 5: chatbots and phone bots
Every day, insurance companies contend with a constantly high volume of enquiries that reaches them through a variety of channels. AI-based systems help insurers respond to customer questions and requests faster, around the clock and without long waits. At the same time, they help insurers automate processes, relieve the burden on staff and reduce costs – without negatively affecting quality.

Trend 6: fraud detection
The German insurance industry faces the challenge of identifying and stopping fraudulent insurance claims. Every year, over EUR 50 billion in compensation is paid out in the non-life and accident insurance sectors alone. According to estimates, at least 10 % of those payments are attributable to fraudulent behaviour. Active fraud detection therefore plays an integral role in the initial processing of claims, especially when those claims involve damages related to quantity. The focus here is on detecting new fraud patterns. Approaches based on AI will enable insurance companies to quickly detect new patterns of abuse and take appropriate countermeasures. An AI-based fraud system should be seen as a learning system that supports claims experts in uncovering potential cases of fraud.

“After using generic approaches, the insurance industry is transitioning to combining a mix of modern analytics for internal and external data with targeted algorithms for specific applications,” says Stefan Riedel, member of the Executive Board in charge of the Insurance business division at the IT service provider adesso SE. “This increasingly enables insurers to go beyond ex-post adjustment and generating prices from models that use past data, instead allowing them to be more and more part of a prevention strategy for their customers. The result is value-added services and offerings that meet clearly defined needs of policyholders.”

Stefan Riedel

Stefan Riedel is responsible for the Insurance division on the Executive Board of adesso SE. (Copyright: Martin Steffen Fotografie)

Do you have any questions?

Do you have questions about our press service or a specific request? Please contact me.

Learn more

Save this page. Remove this page.