6 key differences between a behavior-based FRAML solution and rules-based systems
As a part of Verafin's marketing team, I’m often asked to explain how a behavior-based fraud and AML (FRAML) solution differs from software that focuses either on anti-money laundering rules or fraud detection rules. There's a lot of confusion out there not only about the differences, but what these rules are. After all, not all vendors out there call themselves rules-based. Here's a simple test: if the software involves 'if/then' statements or you have to define the dollar value and time frame of the activity you want to catch, then you're looking at a rules-based system.
So to help create some clarity I chatted up some members of our analytics team to get a better understanding of what the big differences are. They know their stuff, but as computer engineers they tend to be a bit technical. To keep it simple, I've broken it down into a quick list of six of the biggest contrasts.
1. Consolidated vs Siloed
FRAML brings everything together into one solution to effectively monitor and manage AML, anti-fraud and compliance. Software based on rules or other older generation technologies tend to focus on a narrow range of transactions. The data sources and detection are siloed. A behavior-based solution analyzes all data sources together without making analytical compromises. It gives a more complete picture.
2. Intuitive vs Confusing
Good. Bad. Normal. Suspicious. These are behaviors that we all understand. And because you understand the concept of behavior, it’s just easier to understand what you’re supposed to do. Unlike rules-based software, where you have to manage thousands of rules for money laundering and fraud (resulting in duplicate alerts!), you can configure a behavior-based system to look for unusual activity.
3. Next gen vs First generation
Rules were great when they were created – in the 1970s. But today the simple combinations of 'if/then' statements just cannot effectively detect suspicious activity that falls outside of set parameters. With new fraud schemes emerging almost daily, and money launderers savvy enough to know the limits of rules and find ways to work around them, new technologies are needed to stay ahead. Next generation solutions use a combination of fuzzy logic, profiling and networks to determine if behaviors match those that are expected or if they're indicative of money laundering or fraud.
4. Flexible vs Rigid
Which would you rather do? Tune the customer or tune the system? Or to put it even more simply, work from your understanding of how you expect your customers to behave, or spend your time trying to figure out the best rule combinations to help you? A rules-based system means endless cycles of turning on and off rules in an effort to reduce them to a manageable number where you can create a manageable workload while hoping you’re not missing anything.
5. Know your customer vs Know your rules
Similarly, it’s a lot easier for you to determine if unusual behavior is legitimate, and adjust behaviors accordingly. You can think about behaviors, instead of rules. Let's say Susan Potter is a customer of yours. Susan is a letter carrier, who has held an account with your financial institution for about six months. One day you get an alert from your behavior-based system telling you that Susan is depositing cash and checks much higher than her normal activity - sometimes her cash deposits are four and five thousand dollars. Worried that this might be structuring, you investigate. What you find is that Susan recently started selling paintings in a heavily populated tourist area. Susan's normal transactional behavior has changed, and to match that, you change it in your software. With a rules-based system, you would continue to see alerts for Susan, and she would become part of your growing list of false positive alerts to ignore.
6. Endless possibilities vs the end of its evolution
There’s no limit to where you can go with a behavior-based solution. Adding new customer data is simple, and by incorporating various statistical databases, the system can recommend more accurate settings. On the flip side, rules-based systems have reached their limt. It’s more about rule management and pruning these days in an attempt to reduce false alerts.
Of course, this is just a very high level look at the differences. If you're using software based on rules, why not share your war stories? Let’s see if we can get a conversation going to help people understand the differences.
