We consider Investigation and Counter Fraud to be an integral part of our day-to-day operations, strongly intertwined through the very core of our business. Our instructions are not just screened, but analysed by Accredited Counter Fraud professionals and experienced investigators, who are consistently and constantly involved in the entire life-cycle of our handling processes. To us, investigation is not a bolt-on or stand-alone service; it’s integrated into the heart of our business.
We spoke to our Head of Counter Fraud Brendan Gillooly ahead of his speaker session at CILA 2018 about how the claims industry is evolving to combat fraud more effectively. Can AI replace human intervention? Are instant quote tools encouraging fraudulent behaviour? In the digital age when proof of purchase can so easily be faked, what can be considered trustworthy claims validation?
Where do you feel the insurance industry is addressing fraud and its detection?
This is a really complicated question. As a broad response, I think we need to be honest and say we are behind the times. I witness too often an approach that is still reactive to fraud, rather than fully proactive. The market isn’t looking ahead to how fraud is evolving, spotting its potential or the trends ahead of it actually occurring. This said, I am seeing signs of a change with some insurers reassessing their approach.
With advances in technologies and a move towards automated processes with inbuilt intelligence, the resulting complex algorithmic logic created is enabling stronger, more robust anti-fraud indicators; such that they do not replace the human interaction, rather complement it, signposting to those assessing claims the possibilities of fraud, creating more capable escalation processes.
Do you see this change as a revolution or slower evolution?
Clearly, there is a move towards claims being handled more online and with lesser human touch points. However, this is by its nature being introduced cautiously, so is currently happening in pockets, rather than across the industry as a whole. Obviously, there will be a tipping point at some juncture, but for the foreseeable future I would have to see these changes as a slowly-evolving evolution.
How is this affecting your strategies?
Slow as it is, there are changes being made and we need to adapt our thinking to meet them. As already said, currently the vast majority of fraud detection is post-event and managed by human interaction in the lifecycle of the claim. We are now having to consider pre-event markers that identify the possibilities and perhaps probabilities of fraud activities that post-event would be used to signpost towards fraud yet, in the here and now, are used to trigger human intervention.
Even at this stage, however, we have become aware of a potentially worrying trend where we have seen those relying on automated processes adopting a tick-box mentality to reviewing potential fraud markers, and in some cases ignoring the full data in front of them, which clearly show other indicators of actual fraudulent activity. We would caution, therefore, that these skills must not be forgotten or overlooked and that in building new strategies, all the right checks and balances are put in place.
It seems that you are saying that we now need to move towards a position of really understanding the potential fraud markers at the point of sale?
I am indeed, and, to be fair to the industry, this is something that has been recognised for some time. Some good investments have been made to create more sophistication in building investments that are supported by organisations, such as the Insurance Fraud Bureau (IFB) and its data submission platforms for information sharing.
Even better that this is being supported by other bodies such as the ABI, who is mandating that its members support this sharing. When we describe data, this isn’t just active risk data, but extends out to quotes and the like across a broad range of risk classes. Having this vast database available has become a massively powerful tool for us all. A tool that extends to show trends and all manners of discrepancies, all of which are tracked by the IFB and then reported back via their systems of alerts and regular updates.
Hopefully the slower moving claims equivalents can learn lessons from this and quicken its pace towards proactivity.
Thinking a bit more about the pre-purchase activities, there has been a move to simplify and game-ify the buying process, encouraging users to “play” with their quote. Notwithstanding the proliferation of pay-on-use and more esoteric approaches to risk coverages. Do you see this driving an attitudinal change amongst users where it will become more difficult to spot “odd” behaviours?
To a degree, it will change how we need to look at things. The simple truth is those buying policies come onto these sites knowing what they need to insure and with better user experience will be better directed to what they need to buy. If they start playing with things like security arrangements, sums-insured, and the like, this may more easily flag contra-indicative behaviours, which should actually make things easier to spot.
The short term, time on risk style policies are more difficult to consider and really is an area that we think the industry needs to strongly consider in detail, as I am not sure the fraud elements and opportunities have been properly thought through.
Clearly you have shown a number of areas where fraud and its detection have come to fore in the industry’s thinking. Do you feel this has helped create a culture where insurance fraud is seen for what it is, a crime, rather than something that many see as “fair-game”?
Traditionally, it is fair to say that fraud investigations have been tied to the materiality of the sums involved. We have seen this trend change with some insurers now taking a clearer moral stance to fraud, irrespective of size, actively pursuing those committing it, bringing them to justice and using things like the IFB database to ensure they cannot simply move on to the next company to victimise. Sadly, this is not happening across the board and there is still a way to go.
Why do you think this is the case?
I think it is mainly due to the word fraud itself being so emotive. The minute this becomes flagged, naturally everything becomes adversarial along with the causes and consequences this brings to all concerned. We always use adapted language and keep the “F word” out of the vocabulary as far as possible. To us, it is about being tasked with undertaking further investigations to verify facts and establish liability engaging. It is then only at the conclusion of this that the certainty of whether a fraudulent act has been committed becomes apparent and, armed with the facts, appropriate action is taken.
“To us, investigation is not a bolt-on or stand-alone service, but is integrated into the heart of our business.”
Finally, you’ve mentioned recently about how easy it is to manipulate web pages and online profiles to show whatever you want. Could you tell us more?
Proof of ownership has always been a key part of any claims validation. In old, analogue days, this was about physically supplying receipts. Now, with the massive growth of online shopping, it can be about sending over PDFs of web pages demonstrating purchase, delivery and cost. Without going into details of how, the actual source code of those pages can be changed in seconds for them to become whatever is required; with the resulting documentation then appearing to be entirely valid. This is therefore so difficult to spot and, unless counter-checked, will too often pass as authentic. We are seeing this starting to emerge too often, and not just with receipts. Entire threads of email conversations are likewise being altered. Being aware of it, we look for it, but again this just serves to demonstrate how technology is marching ahead at a crazy pace and how we as an industry need to adapt at the same speed.
To give some comfort, within the boundaries of respecting others’ privacy and regulated activities governed by the GDPR legislation, there are a variety of ways in which we are able to access various data about individuals and their online footprint, including where appropriate, without express permissions, to validate and verify enquiries. This has often led to actionable evidence that can be gathered and used to prove fraud. In the modern world, social media has become an investigation tool all of its own.
Brendan is speaking at the CILA Conference on Thursday 20th October and would be happy to discuss more on counter fraud if of interest to anyone attending or reading this. Get in touch with him and find out more about our fraud investigation services here.
Follow us on LinkedIn where we’ll be adding the insights from the CILA Conference.