I visit with a lot of BSA and Risk officers in my role as sales engineer for our financial crimes solutions. And let me tell you something … I hear a lot about false positives. I have heard folks in our industry call these the bane of their existence. Whether it’s with sanctions screenings, AML, card processing or any other type of fraud, getting too many false positives is a bigger risk for your institution than you may realize.
I found one of the insights on false positives from a recent fraud study conducted by SAS on mid-market banks ($10B-$50B) to be particularly compelling:
About 40% may be reporting a false-positive rate higher than 25 to 1. (In other words, 95% of transactions that are legitimate are predicted to be potential fraud).
Imagine that – if your investigative staff is seeing 100 alerts, 95 of them are false positives.
That is astounding. This also makes me wonder how it could affect the morale of the team. Are you able to keep your fraud and AML investigators engaged and committed to finding real problems even when they see so many false positives? Does your investigative staff longevity match the other operational areas of your FI? People who come to work for your FI want to make a contribution. The type of person drawn to an investigative role is especially tuned that way. Constant false positives could naturally erode the level of job satisfaction.
I suggest that the answer isn’t to ignore the problem and pretend it doesn’t exist, or to hire an army of investigators to filter through the white noise. It’s to address the root cause of the issue that is triggering so many of these false alerts to begin with.
So what can you do? Let’s start at the beginning. Are you currently able to quantify your false positive rate? We should be measuring this both on the AML and fraud sides of the bank. As you likely know, they are both important in different ways. If you don’t currently have a way to generate reports that tell you how many of your system-generated alerts result in cases and then SARs, that is a big problem. Understanding this is job number one. Begin by sitting down with your top fraud and AML investigators and asking them how they think your solution is working. What could be improved? Where are they seeing a frustrating volume of false positives?
Once you have proven that you can quantify efficacy of your alerts, job number two begins with refining, testing, and monitoring your progress in reducing the number of false positives. Come back to your team and listen to them for signs of progress. This requires patience and dedication. You must ensure that the triggering events for alerts are supported during investigation by a holistic risk view of your customer.
Take the time to quantify and tune your models and rules so that you know that you are covered. Perhaps you do need more investigators. Maybe you lack automation and have too many manual processes. Maybe correlation, automation, or integration are what you need. Place measurements on how effective your models are and determine if you are getting a result that would be expected from a bank with your geography and risk profile. Ensure that the solutions you are using are helping you achieve this by correlating the data you have and allowing you to measure, test, and refine your output. Using these guidelines, you can save money and improve morale. Is it time for your FI to tackle the false positive problem?