Five Reasons Your AI Project Never Made It To Market
Many people enjoy watching a Netflix show or reading a light novel to take their minds off things. Me? I read postmortems of failed AI products.
I’m not the only one obsessed with figuring out why so many AI companies never achieve the success they aim for. Analysts and AI experts have tried to assign numbers to this phenomenon. Supposedly, only around 1 in 10 data science projects make it into production.
If I were superstitious, I’d say there’s some kind of curse sending AI products to an early grave. But I’m not — I’m a data scientist. Having worked in AI at Uber and now as the CTO of Evisort, a startup automating contract management, I know what it takes to launch AI products successfully. I’ve also seen how it goes wrong.