Creating value for life insurers

Interestingly, according to The European Union Artificial Intelligence Act expert systems such as automated underwriting engines are considered AI. The insurance industry has been using these sorts of expert systems for decades to support medical underwriting. More recently, there is also a growing trend in the use of automated claims engines to improve process efficiency for claims processing. For example, Hannover Re’s Claims Assessment Module (CAM) is an automated solution for declaring and assessing insurance claims. Using decision tree logic, it simplifies the task of processing a claim for insurers and their customers. This saves time and resources for the insurer, and provides a simple, professional and positive experience for the end customer.

In the past decade or so, machine learning has also become more popular to support insurers with modelling for specific tasks. Machine learning models are trained for specific tasks, and typically require large quantities of training data. In an insurance context, there has been a rise in the use of machine learning techniques to predict insurance risks such as policyholder lapsation or claims risk. This sort of technique can also be applied to help improve medical underwriting, for example by enabling deeper multi-variate analysis of the data for manually underwritten cases. For example, to identify cases with a high standard rate and a very low decline rate that can be transformed into new underwriting rules. This sort of enables can help to enable higher straight-through-processing rates for the insurer, thereby reducing the amount of time and cost the insurer spends on manual underwriting and speeding up the underwriting process for the customer.

There are also examples of the use of AI on the claims side to help identify document fraud. Document fraud is costly to insurers, with the insurance industry facing billions of dollars of losses due to false claims and identity theft. The burden of proof is on the insurer to demonstrate the fraud, and this can be a very time-intensive process. Moreover, with the growing accessibility of AI, fraud is becoming more sophisticated and is not always visible to the human eye. Using AI tools to help identify claims fraud can therefore be highly valuable to the insurer. Hannover Re Life Australia has partnered with the Australian technology company, Fortiro to use AI for document fraud checks when documents are submitted at the point of claim for disability insurance. Fortiro’s fraud protection solution, Fortiro Protect, conducts a content analysis, a layout and visual analysis, and a digital footprint analysis for submitted documents. If Fortiro Protect flags a high risk of fraud for the documents, the claims assessor can follow up and investigate suspicious claims in more detail.

Emerging opportunities for GenAI

Most recently, generative AI has exploded in popularity. Generative AI can generate new text or images and can be applied to a range of tasks. Due to the models having been pre-trained on vast quantities of data, large volumes of additional data are not typically required by users when using the models. In an insurance context, there is potential for Generative AI to support process efficiency improvements and enhance customer experience. For example, chatbots may be developed to support agents and insurers in providing relevant and timely information to their policyholders at point of sale as well as throughout the lifetime of the policy. In terms of medical underwriting and claims management processes, Generative AI may also assist in terms of summarising medical documentation received at point of underwriting or claim. This can help to speed up manual underwriting and claims assessment. For now, the use of GenAI in insurance is still very early-stage. Some best practices when exploring the value of this technology is to keep a human-in-the-loop to verify the accuracy of the model’s outcome as well as to be aware of potential biases within the model and its training data.