As fast as Big Data tools and techniques improve, fraudsters are still keeping one step ahead and managing to illicitly obtain corporate information and funds, warns KPMG. To counter the threat, KPMG's UK and EMA Leader for Data & Analytics calls on companies to move towards a more analytical workforce, arguing that doing so will make it significantly easier to spot the anomalies which can often indicate fraud or improper conduct – and ultimately catch the fraudsters.
According to Eddie Short, it is possible to analyse Big Data to find the 'outliers' that can identify fraud. He highlights a variety of tools which can run complex algorithms to identify abnormal behaviours and argues that these can be 'tweaked' to identify rogue traders or other types of malpractice.
Eddie Short comments: "Fraud and corruption usually occurs at an individual level, so it's very hard to spot using Big Data alone. Companies need to switch to a more analytical workforce to increase the likelihood of eyeballing those anomalies which exist as a result of fraudulent activity. Many organisations have some of the brightest people on the planet – hackers and crime bosses – working against them. It's a metaphorical war of intelligence!"
According to Short, by using Big Data to fight fraud, organisations need to:
- allow time for analysis: enabling their employees to test, experiment with and interpret Big Data, without fear of failure, to gain genuine business knowledge and understanding, so they can identify best practice and fraudulent activity
- capitalise on present day software: using the latest analytical tools to enhance the value of information and knowledge from Big Data to help identify bribery, corruption and money-laundering
- look to the future: by creating more sophisticated algorithms to enable the processing of even greater amounts of data, faster and more accurately. This means less opportunity to hide bad or fraudulent practice. This will also include greater use of cognitive and Artificial Intelligence systems which can be much more predictive.
Hitesh Patel, UK Forensic Partner at KPMG, adds: "The sheer amount of data created on a daily basis is being used as a smokescreen by individuals intent on defrauding organisations. They use the mass of information that exists to bury what they are doing, but the key to beating them revolves around analytics using forensic techniques for greater transparency and having fewer places to hide.
"Analytics alone is not the magic bullet but it is certainly the fastest way to locate the bad apples in a barrel. It's a powerful technique to highlight oddities that require wider investigation, but real success will only happen when the capabilities of analytics tools are merged with those of skilled staff with a forensic eye, able to use these tools to spot odd transactions before they become a costly trend."
Eddie Short concludes: "15 years ago, we were watching IBM's Deep Blue beating chess Grand Masters by analysing moves and outcomes. Now we're seeing systems like IBM's Watson Technology beating the champion contestants on the US show "Jeopardy" because it anticipated and analysed the potential questions. It's one thing to work out what the answer is; the skill is working out what the question to the answer is. It's the same with fraud: you need to be anticipating the right questions and that requires the right Big Data tools and people to be able to achieve it."