It’s not uncommon for applicants flagged as potential frauds to be processed without being manually underwritten. Why? Simply because it is unrealistic to expect 100% of all referrals to be manually worked through. The sheer extent of referrals means that underwriters are suffering from capacity fatigue and a major backlog.
What does this mean? Ultimately there is a higher level of fraud passing through the systems leaving you as a lender exposed to high potential losses and at a much greater risk of growing fraud. The level of referrals being manually underwritten, when obviously not fraudulent, are consuming the underwriters’ time and costing a lot of unnecessary expense too.
More importantly, it can mean you have a constant struggle to adjust your referral rules in line with the ever evolving trends in fraud. This can impact the predictability of fraud referral rules over time meaning your overall strategy performance also deteriorates over time.
The result; it is all a challenging balancing act. This is mainly derived from the common inability to balance the capacity of manual underwriting with the volume of fraudulent cases presumed. Having the ability to maximise the volume of referrals with the amount of fraud captured, with minimum effort, is vital. But, it’s not always that simple. If it was, a process which considers this – and solves it – would be in place.
Work lists are already common practice. The pros; they can help apply a priority to referrals – in order of risk. They can allow matching of referrals with the specific underwriting specialists meaning you can segment them into areas of expertise which would save time and make the most of the resources at hand. Furthermore, they can be grouped by fraud rules – helping you consider and understand which aspects of fraud is already a likely risk for example. However, again, it’s not that simple.
Individuals cannot necessarily order the risk ranking independently. There is no indication of when to start working on another list and there is no consideration to a blended rule view when assigning the work task.
What if you could ‘score’ each referral? Scoring could complement referral work lists by providing a single reference point for the probability of fraud associated with each application and managing the working orders accurately – applying an automated view of the highest risk of fraud per referral. This could enable you to capture the same amount of fraud, but on a fraction of the referrals being manually worked. Scoring considers the multiple arrays of fraud rules too – which considers and automates this risk likelihood. Using insight and trends you can then refresh the score to reflect fraud which is based on real time trends and fraudulent activity.