Potential remains unused

Almost two thirds of the insurance companies surveyed by Bearingpoint in Germany have not yet deployed any productively usable AI services in process automation. In fact, many companies have only just begun to gain experimental experience with AI technology.

According to Bearingpoint, many companies are still struggling to identify and detail use cases. The development of AI expertise is still significantly slower and in many cases the specialist department is unable to recognize the potential for using AI services.

For example, 43% of respondents cited an unclear technical demarcation or definition of the use case as a challenge. The study therefore urgently recommends building up AI expertise in the specialist departments and in the supporting IT organization.

The right data for the desired AI services is also often lacking or is not available in the necessary data quality, as Bearingpoint emphasizes. For example, 42 percent of those surveyed stated that the existing data is unsuitable in terms of quality and quantity.

According to the study, the required data should therefore be collected and provided in sufficient structural and content quality in advance of the use of AI services. The survey also revealed that 70 percent of insurance companies have so far only used their own data for the development of AI services. However, the use of in-house data is increasingly reaching its limits.

Sufficiently qualified and quantified data for high-quality AI services is often not available. According to Bearingpoint, it is always advisable to "buy in" external data in order to close data and information gaps and thus drive the development of AI services forward.