So many of us work with anti-fraud and other systems that generate alerts and work queues. In general, we hope they will find the rogue event that makes all the difference.

Maybe it’s like shooting fish in a barrel, but in most cases, our specialists spend a significant proportion of time wading through anomalies caused by poor data matching or process triggers that are just too sensitive.

So, why not ask those analysts about the worst part of their day. I bet sorting through the false positives is near the top of the list. And I bet the same scenarios keep presenting themselves. Now that would be annoying.

Now ask yourself when your business did its last review of its anti-fraud operation. We are all aware of the InsureTech noise and buzzwords like machine learning. There is no doubt that the level of computing power and smart systems have accelerated in recent times, but it is knowing how to apply that technology to your business.

What if that repetitive process of removing false positives could be “learned” by the machine?

I saw exactly that recently and that would be my challenge to all the existing providers in this space; What are you doing to apply machine learning to your system outputs?

For insurers, you need to review your fraud referral criteria. For vendors, you need to review your systems with the eyes of a fraud practitioner.

Always listening: Please let me know if you recognise any of these issues and need some advice on maximising your data. Always interested: it’s always good to learn about new and exciting technology.