I recently attended a fantastic roundtable put on by Financial Times in New York and as I’m sitting listening to a group of folks that specialize in anti-money laundering and fraud, one person stated that people detected the vast majority of fraud or money laundering with tools playing a supporting role.

By itself, this seems to be a bit damning toward the technical sector essentially stating that they aren’t any good at detecting fraud. Or at least their tools aren’t any good. But technology has always played a catch-up role when compared to human intuition. It could be simple things like highlighting the right statistical inconsistencies for analysts or complex things like playing chess against the world’s best, but we’re all still trying to mimic a human (with scale) by building intelligence into systems.

Stop!, by Qfamily

So this really begs the question, whether you are in the information security business or the fraud prevention business, where is your first line of defense? I’m willing to bet it falls in line with the observations from the roundtable and it is entirely human focused. But I’m also willing to bet that your company realizes this plan isn’t scalable and is trying to find ways to build human intelligence artificially into our infrastructure to aid the humans. For example, humans cannot read millions of logs manually, they have systems that triage mountains of work into molehills of actions for further analysis.

So the question becomes, how do we accelerate this so we can get to the point of more front-line defenses being built into artificial intelligence instead of relying almost solely on human intuition?

Companies require comprehensive visibility into events into their network with the ability to incorporate both internal and external sources of intelligence to create actionable intelligence that can feed an automated, agile control set. Sure, it sounds a little like the beginnings of Skynet, but those that fear the rise of the machines may choose to build the intelligence without the capacity to act, thus still requiring human interaction but theoretically with better information. The goal still needs to be furthering our ability to transfer human intelligence into systems to help us do more with less (and reliably!).

This post originally appeared on BrandenWilliams.com.