The cost of living is rising fast. Official inflation rates vary, but the reality is being felt by people around the world as food, fuel, energy and other essentials are increasing by anything from 20% to almost 100%.
That will inevitably lead to increased attempts at fraud, exaggerated claims, ID manipulation and more. BAE Systems Digital Intelligence takes a look at the cyber & fraud risk landscape and how insurers can navigate through it.
According to the government *, some 39% of UK businesses had reported a cyber attack in the previous 12 months. Typically the cost of dealing with each incident, for larger businesses, was around £19,400. The report also flagged that around 16% of businesses reported using versions of MS Windows older than 8.1. [*more detail here ]
But cyber attacks have been morphing over recent years, from Denial of Service (DDOS) plus ransom, or hacking of data, into attacks on physical assets, such as pipelines, energy distribution or key buildings. For example a Lloyd’s report in June 2022 looked at how vulnerable modern systems are to third party control.
For insurance brands, problems during the inflationary economic cycle are likely to be:
• Exaggerated or fake claims
• Fake ID or false information at point of quote
• Organised multiple claims via different insurers
LINKING THE DATA TRAIL
One feature that has emerged in the last few years is that several criminal activities can be linked to one gang. So for example, phishing, crypto scams, people trafficking and car theft can all be part of the same organisation’s daily operations. The clues are often to be found when you drill down into the data.
They can be anything from multiple shell companies, or bank accounts, registered to linked addresses or surnames, repeated attempts at making travel sickness, or crash-for-cash claims to different insurance brands and so on.
Given the pressure on household budgets this year, it will not be unusual to see lots of policyholders change their addresses, as they search for more remote jobs which allow working from home for example, or tenants seeking to exit the high rental market of London and the south east.
This is where insurers need to automate the checking of data that is being manually inputted at the point of quote – does it match what is on file, or available on shared data bases like the DVLA, MIB or Electoral Roll?
CLAIMS ANALYTICS
As claims become ever more automated, it also becomes more important to check the ID of the person submitting statements or images. This involves cross-matching IP addresses or devices being used, looking at the time of day or night when data was uploaded, or checking bank accounts or payment services being used match prior claims logged.
All these clues can be built up to create a true 360 degree view of the claim: who is involved, their address, vehicle and claims history and so on. Don’t forget that a correlated claim may be a follow-up to a house or injury claim made by the same person, or someone linked to that person, within the past 5 years.
With an expensive winter ahead in terms of energy bills, and no end in sight to high fuel costs, insurers can expect more attempts to obtain `cashback’ via fake claims or exaggerated costs associated with genuine incidents.
The best prevention is analysing data for typical red flags that are potential fraud, or cyber attack, indicators. Is the claimant in arrears with rent, mortgage or card payments; have they recently changed or lost jobs?
But what about novel fraudulent or cyber attack behaviour, for which there is no typical red flags because the behaviour is new and unexpected? In this case machine learning based anomaly spotting can be used to support detection and prevention.
By joining the dots online insurers can build an index of data which acts as a risk score. It might not be that every high score is a high risk, but it gives you the opportunity to analyse in more depth before proceeding. You can read more on how BAE helped one insurance brand to combat fraud in this case study.
by David Nicholson, Senior Data Scientist, BAE Systems Digital Intelligence
Be the first to comment