Reduce Fraud? Use EDM...

Interesting article in Insurance Journal ABI Calls for "Radical Action" to Reduce U.K. Fraud talking about insurance fraud costing each family in the UK $1,000 annually. The article talks about some steps that could be taken to improve this situation.

What the article does not do is talk about how software, especially decisioning software suitable for an Enterprise Decision Management approach, can address this. Other industries, like credit for instance, have had great success in reducing fraud using business rules (to target spot fraud or new fraud trends) and analytic models that predict fraud (like neural nets to track out of pattern behavior). One example here is Fair Isaac's product Falcon Fraud Manager. Similarly the same combination works really well in claims fraud here in the US (when there is a political will to use it) - Vericomp Fraud Manager is our product there. This stuff works - it just needs to be applied to the industry.

So what would I suggest? Well I might suggest that the ABI:

• Get insurers to pool data to build effective fraud detection models using neural network or similar technology
• Work with software companies to develop the kinds of software products that can leverage these pooled models and business rules to help detect and deal with fraud
• Then do the other things they listed like push the government to spend more on enforcement etc.