The sheer size of data collected by any large business has created a digital undergrowth where fraudsters can hide. But the technology criminals use to commit fraud can be successfully turned against them.
KPMG, the firm of auditors and business advisers, has some 400 UK staff employed in forensic investigations for clients concerned with real or suspected fraud. Hitesh Patel, a forensic partner with the company, acknowledges that, at the first impression, technology appears to have raised the bar for investigators.
“We create so much information on a daily basis that it is tough for organisations to see where risks are being created,” he says. “This does make it easier to hide improper activity.”
There are software tools designed to hack through the dense mass of data and spot the seemingly random correlations that indicate wrongdoing. KPMG makes extensive use of these products, employing complex algorithms to identify abnormal behaviour.
Featurespace is typical of the companies supplying such tools. Describing its work as behavioural analytics, Featurespace is backed by Mike Lynch, founder of the Autonomy data analysis company, among others.
His involvement is no random investment. Featurespace springs from the same school of mathematical thought applied to real world problems that has seen Cambridge mathematicians, such as Dr Lynch, setting up a string of analytical software houses. In short order, Featurespace can be calibrated by its users to spot rogue traders or other fraudsters by isolating patterns of data that are not correct.
Mr Patel recounts how much of his work involves picking a chunk of data representing a week or month out of one year. This is then scrutinised through the lens of an analytical tool that will highlight potential discrepancies in employee conduct.
There are software tools designed to hack through the dense mass of data and spot the seemingly random correlations that indicate wrongdoing
And the broader economic context is one that is fuelling fraud, he says. “Some people turn to fraud to maintain a lifestyle in tougher times. And as companies restructure and change their operating model this can create opportunities for fraud.”
The technical tools that KPMG employ narrow down the field so Mr Patel and his team can bring their experience to bear. Various indicators, such as invoices being processed over Christmas holidays or payments being made to suppliers in advance of normal terms, are brought up by the analytical tools. These are transaction tests that can be set up before the smart software is unleashed.
But however much a fraud-savvy business invests in clever computer tools, the real skill still lies with human beings. “Just buying the technology is not enough,” says Mr Patel, “you need a forensic mind to get behind it and drive it.”
One of the points where software analysis of company data can fail to detect fraud concerns linkage. It’s very easy to run a tool over the accounts while ignoring information from the personnel department that gives a clue to a rule-breaking mentality.
“People tasked with monitoring for fraud need to look at everything. So if the human resources people are getting complaints about one manager bullying staff – that should be seen as more than a legal matter. It could indicate an underlying tendency to ignore company policy and legal constraints,” says Mr Patel.
Using the sharpest software produced by the finest minds is certainly evening the odds in favour of honest practice. But that is no cause for complacency. Every business needs to embed a fraud detection mindset right across the board in order to thwart malpractice and theft.
