How AI Is Tackling Insurance Fraud

Insurance is one industry where digital technology arrived late but is having a significant impact. Away from the “insurtech” boom, it’s creating a cat-and-mouse game between fraudsters and those who want to catch them.  Our team’s insurance expert Laurence Smith explores how the industry is embracing digital tools such as voice and emotion recognition to combat insurance fraud.

So?

£1.3 billion is an eye-watering sum to most of us.  Even in an era when companies like Apple. Microsoft and Amazon are hitting trillion-dollar valuations. That’s the figure the Association of British Insurers (ABI) reports that insurance fraud costs the sector every single year, despite the 125,000 dishonest claims already detected. It’s the reason why insurers are investing £200 million each year to identify fraud.

Insurance fraud is real and happening right now. Organised criminal gangs are setting up “crash for cash” style motoring scams.  However, the problem is not just driven by a reckless minority - the source may be closer than you think. The guy next door might be guilty of recklessly supplying false information, and many otherwise law-abiding citizens seem to regard “faceless” insurance companies as fair game.  

The little white lies people tell when claiming mount up to significant payouts and results in a hike in the cost of premiums for honest policyholders. Insurance fraud is not victimless. The normalisation of fraudulent behaviour is socially corrosive and erodes trust.

The truth is, insurance fraud doesn’t discriminate. It cuts across every type of insurance and affects society as a whole.

So What?

Anyone who’s ever received an email that starts with “I need your bank account details so I can transfer money I’ve just inherited from an exiled prince” can attest that fraudsters thrive in the online world.  They can set up multiple IDs, test numerous scenarios until they identify those which are acceptable to insurance algorithms, and submit fake documents during verification checks. 

There’s also been a rise in so-called “ghost brokers”. Fraudulent insurance policies sold by illegal insurance advisers, carrying the logos of genuine insurance companies, sold on to the end consumer.  

As fraudsters grow more sophisticated, traditional methods used by insurers to root out fraud are increasingly ineffectual.  The challenge is exacerbated by the demise of the high street broker, who in bygone days was at least able to raise suspicions by looking a customer in the eye. 

Ironically, the digitisation of business models – which, in the short-term, has increased the opportunities for fraud – also provides a vital part of the solution.  Here are some solutions being used;

Triangulation of data from multiple sources, including data bureaux and public records, can pinpoint anomalies – in particular, any mismatch between the data supplied in an application and the data held on file.  Technology is assisting with cross-industry anti-fraud collaboration.

Device tracking enables firms to highlight multiple transactions from a single source (tablet, phone or laptop), regardless of factors that can be manipulated such as the connection used or the user credentials.   

Pioneering analytics can follow activities from a suspected fraudster across multiple, seemingly-unrelated accounts.   And, when phone calls are involved, call centre AI is being programmed to flag unusual behaviour which can manifest in the most innocuous ways – vocabulary, intonation, voice patterns and other emotional cues.  

Modern Telematic systems - In-vehicle technologies like telematics systems and dashcams are increasingly helping tackle motor fraud.

Insurers have partly acquiesced in this trend through their past reluctance to prosecute known fraudsters.  This has meant that, for criminals weighing up the payoff versus the downside of different schemes, insurance fraud has generally been a low-risk option.  

Recent months have shown more appetite by insurers to initiate legal action, boosted by the greater willingness of judges to impose meaningful sentences, and also by the growing brazenness of the scams.  

So Now What?

Insurance - like cybersecurity - has become a tech battleground. If your industry is being changed by technology, here are things you can do:

Understand. What can insurance teach you about your sector? Can healthcare, for example harness the emotional analysis being used by insurance call centres? Or the algorithms used to track vehicle journeys?

Think. How will digital tech transform your industry? What is the equivalent of this battleground for legal, agriculture or manufacturing?

Apply. Is there a revenue loss in your industry which could be plugged by using similar technologies?

The £1,155,789.38 life insurance claim lodged with three insurers – Aviva, Scottish Provident and Legal – by a grieving widow whose husband had “died of brain fever” on an overseas business trip was somewhat undermined by his reappearance in the UK under an assumed name.   

Be ready for many more stories in this vein as AI is coupled with big data … and targeted like a searing laser beam upon the dark entrails of that £1.3 billion monster.