Fraud detection with machine learning leads to £32 million saved on claims payouts for UK firm
02 November 2018
CASE STUDIES
What if Machine Learning could save your insurance company an extra £32 million on claims payouts per year, simply because of its accurate fraud detection ability?
With over 600 litigators handling over 100,000 insurance cases each year, our client was keen to find out whether DataSparQ could create a Machine Learning solution to increase their litigator’s processing power and improve their fraud detection rates.
Our client is a UK firm responsible for the legal approval of insurance claim payouts on behalf of over 70 UK insurers, including four of the top ten motor insurers. The insurers pre-select their most complex cases and our client’s 600 litigators manually process each claim to ascertain whether it is fraudulent. Of the 100,000 claims processed each year, roughly 7% (6,800 cases) were suspected to be fraudulent. That equates to a saving of £225M in claims payouts.
Machine learning proves its worth
DataSparQ’s data scientists created an automated machine learning solution to help the claims processing team to efficiently and effectively detect fraudulent claims and prevent payouts on them.
The solution is able to identify those cases that are legitimate claims. This equates to 60% of cases passed to our client, and has therefore reduced the number of claims having to be investigated to 40% of the original total.
At the same time, the solution is able to identify fraudulent claims more accurately, and of the 40% being investigated, one in two were discovered to be fraudulent. This increased the savings in payouts by a staggering 14%, from £225M to £257M.
What a result!
Because our client’s litigators are investigating 60% less claims, they are able to quadruple the number of cases they take on. And because the Machine Learning solution can detect fraudulent claims with greater accuracy, they now have a tool that is extremely attractive to potential new clients and a major differentiator against their competitors.
Explore another DataSparQ case study for the insurance industry:
We’d love to speak to you more about your challenges and how our data science solutions for the insurance industry can help. Get in touch.