In tech journalism AI is painted to completely evolve industries, replace humans, and all that futuristic jazz. AI lends itself to selling intriguing possibilities in a few years. 

Do we need to go that far? Where are we at present? I've already stumbled on a few new things many companies will probably pay for right now. 

Here's one example. For some who monetise online, there's a specific topic that's sure to frustrate: payment fraud. It became common around the world to the point human analysts - on their own - weren't cutting it anymore. 

Moreover existing rule based systems as well as classic machine learning solutions from the past 1-10 years are either also expensive, struggle to scale, or are too slow to adapt to new fraud patterns in real time. Our current detection systems aren’t advanced enough to detect the newer, and more advanced fraud patterns. 

It's easy to like Fraugster and their practical use of AI. Their AI platform learns from each transaction in real time, combining the thought process of human analysts with the speed of a machine—making decisions in less than a second. With a new self-learning algorithm that mimics human analysts, I think this has the opportunity to analyze the story behind each transaction and say with precision which transactions are fraud and which aren’t.

Because essentially this is an engine that reinvents itself with every new transaction. Powerful. Scalable. Businesses which transact internationally online at volume really need this. We as customers really need this. 

I do want a voice enabled driverless car that dodges a crash, predicts my heart attack, and intelligently arranges a funeral to my preferences... though ... I'm actually more pleased that AI is very capable of solving clear pain points today.