The Use of Artificial Intelligence in Insurance Fraud

Original article was published by Erika Bothma on Artificial Intelligence on Medium


A recent study by The Economist Intelligence Unit (entitled The Road Ahead: Artificial Intelligence and the Future of Financial Services, commissioned by ThoughtSpot) focussed on banks and insurers’ expectation of AI’s impact on their businesses over the next 5 years. The study showed that a wide range of AI technologies had been implemented by banks and insurers alike.

The study’s survey shows the top 5 benefits experienced by financial services organisations as in figure 1 below:

Figure 1: Top 5 benefits from AI adoption (Source: The Economist Intelligence Unit, 2020)

(When consulting The Economist report, note that my figure 1 appears as “figure 2” there.)

With regards to an AI-enabled fraud solution, many of the benefits mentioned in figure 1 are very relevant:

  • Operational costs are reduced by automating the process of checking for fraud.
  • Predictive analytics are enabled by the AI model, based on the insurer’s historic data.
  • Customer satisfaction is enhanced through speedier payment of valid claims.
  • Employee workload is reduced, as the lengthy process of fraud checks is handled by the AI software.

The report also points out that metrics to measure AI success weren’t in place in many instances where they investigated, which is worrying if one considers the high cost of many of these solutions. I’ll give recommendations for measuring AI success with regards to fraud solutions in article 4 of this series.