To be blunt, the universal use of commercial off-the-shelf AML/CFT databases of high risk individuals and entities, which have ruled compliance for the past twenty years, since 9/11, is simply not effective in real-time interdiction of money laundering and the financing of terrorism. The most conservative estimates believe that only a couple of per cent of existing laundering operations are identified, and shut down while in progress, which means the laundrymen are winning; compliance is losing badly in this war.
The reason for the utter failure of the financial world to stop money laundering in its tracks is the generally-accepted principle that, by categorizing bank clients as low-, medium- or high-risk, and thereafter actively monitoring them accordingly, those most likely to seek to move and clean the proceeds of crime are identified. Nothing could be further from there truth.
What do you say about your theory, when a clean bank client, whose background screams legitimate company, with all the bells and whistles to prove it, goes on to successfully move a lot of dirty money through your bank, right under the noses of your compliance team ? Your vaunted databases failed to identify the trade-based money laundering operation that your "good" client accomplished, fooling not only your people, but law enforcement and regulators as well. Your databases failed to catch him; now multiply this one bank client by hundreds of others who have, either in the past, currently, or in the future, beaten your compliance staff, and you understand the nature of the problem.
The solution is not relying upon the risk profile of your bank client, but creating an accurate, and timely, assessment of who has the propensity to commit money laundering offenses in the future, through an effective analysis of information, much of which does not necessarily factor into a simple risk profile. If your program can, with a high degree of accuracy, predict whether or not your clean client will be moving criminal proceeds, then you have a solution that is both effective, as well as efficient. It's not who your client is today, but what you are able to correctly ascertain he might do tomorrow, that counts. Anything less than that is not a winning AML/CFT solution, in my humble opinion.