AI continues to be the buzzword of the moment in VC and start-up land, with nearly every pitch deck having some reference to AI or ML. So how is an investor supposed to make sense of this vast array of AI applications?

The good news is that you aren't. AI has now become such a broad field, with novel research pushing the boundaries of what's possible on so many fronts that no one individual can be an expert in all of them. A recent blog from Julian Togelius, looking to help journalists covering AI, sets this out explicitly. He advises them to always check with experts in multiple fields, and to admit that nobody understands everything that goes on at an AI conference.

So how do I assess start-ups using (or claiming to use) AI? Some of the basics are common across most applications of AI, and checking that start-ups know the basics will catch-out those just trying to jump on the hype. 

The other key for me is to talk about use cases: Who is the target customer and what problem are you solving for them? 

I want to find start-ups making meaningful differences for their customers, not just using some cool new tech for the sake of it.