Poor procurement costs. Procurement Protect is a fraud detection platform that proactively identifies fraud, error, and non-compliance with organisational processes and policies in the purchase-to-pay process.
The platform uses sophisticated analysis to examine data from major ERP systems, identifying anomalies in behaviour and events that describe the building blocks of fraudulent activity.
Procurement fraud accounts for a fifth of all reported fraud according to our Global Economic Crime and Fraud survey 2020, and is notoriously difficult to detect. Fraud, along with innocent error, amounts to a significant hidden cost for many companies.
Even before the pandemic, we saw a 28% rise in the number of organisations in the UK experiencing procurement fraud, and fraud risk exposure may now be even greater with huge numbers of employees working from home and normal processes disrupted. Most organisations may only see the consequences subtly and gradually impact the bottom line – others may not notice any problem at all. It can take years for procurement fraud to come to light – and as we’ve seen in practice, losses can amount to millions of pounds
Until recently, fraud detection has relied on ‘red flags’ that highlight exceptional transactions – creating the need for additional manual checks – but these are well known by seasoned fraudsters and easy to avoid.
We take a different approach, based on our knowledge of how fraudsters work. We identify ‘features’ of possible fraudulent activity – using statistical, rules-based and machine learning algorithms – and then triangulate these to spot subtle patterns in the data that match known types of fraud.
We combine sophisticated data analysis with our extensive experience of investigating fraud to tailor the analysis to each client’s specific circumstances and needs, identifying procurement fraud before it has a chance to escalate.
Procurement Protect integrates quickly with existing ERP systems and can handle large volumes of data, meaning that all transactional data is analysed, not just a sample.