Get Started
Log In
Menu
Get Started
Log In

Risky Business: Dynamic Thresholds and Unheard-of Efficiency with AI

by Kristina Drye, on Jul 22, 2021 12:36:37 PM

The business of screening and vetting can be summed up in one simple concept: A service is offered, and there is a level of risk we are willing to accept when providing that service. Below the level of risk, we aren’t pressed for further investigation. Above that level of risk, we begin to care about liability. A screening and vetting process is the process by which we review populations to sort them into these categories. 

Riskline-1

Sorting like this takes time- especially if you don’t have the tools to help. As professionals, we might be able to get through 100, or even 1000 cases a day (if we’re superheroes). But if our population is larger than that - and most are - we begin looking for a tool to speed up the process. 

Enter artificial intelligence and machine learning. These tools take the manual process from us as professionals and sort cases into the ‘Above’ and ‘Below’ categories with an ease we can never achieve as humans. The tool itself can tell us which portion of our 10,000 entities needs to be looked at because they fall above the risk line and which portion we don’t need to worry about because they fall below the risk line. 

Thresholds

With AI and ML, this distribution is determined both faster and more accurately. These technologies find information we cannot, digging into the deepest corners of the internet to identify the risk we need to keep our organization in compliance. 

This all sounds great- except for one question.

"When the threshold is set for the machine, can it ever be changed?"

Isn’t it dangerous to not be able to change that threshold? What if we decide the risk level is different than we thought - do we, as professionals, actually have control over that? 

The short answer- and the good one- is that yes, you do have control over it- as much as you and your team need or want. The misconception that a risk threshold used by an AI tool is static is luckily just that- a misconception. 

Using GOST, the threshold can always be moved. If you, your team, or your institution decide that the threshold needs to be lowered or raised, it can be - and that can happen as many times as you want. 

There are many perceived gives and takes with employing a new and emerging technology. Trading a dynamic threshold and an inefficient process for a static threshold and an efficient process is not one of them. With GOST, your screening and vetting practice can be both nuanced and effective. 

Topics:Article

Recent Posts