Fraudsters are by nature generally smart and unpredictable- as soon as an insurer identifies a particular scam they change the modus operandi. Aviva states that "Whiplash - fraudulent claims still prevalent".
Great hope is put in AI to help detect and deter fraud. Yet AI is most effective in predictable and repetitive behaviour and human nature.
Many insurers plan to automate claims processes relying on machine learning and AI to combat fraud - especially high volume AD claims. But experience shows that over 30% of home insurance claimants walk-away when faced with having to present text, photo and video evidence in a seamless, sequential digital claims process. AI would have passed these for payment.
At $600 to $700 a pop that's a "load of dosh"
Don't mistake this is a denunciation of AI- it plays an important part in decision making, forecasting, diagnosis and more. That will be my next blog.
At a time when insurance companies are pulling all stops to ensure that customers are sold the right products, a stronger fraud detection by the industry is the need of the hour. In a bid to weed out criminals from the system, insurance companies could also be losing out genuine customers in the process. Insurance still being a push product already needs a strong push for any individual to buy it. Simplification of the process and automation to trigger fraud-sensing at time of underwriting is a good start. What will be closely watched is whether insurers are successful in eliminating country-wide fraud through this process.