Discussing the high failure rates of AI projects
AA: You’ve been in the field for a while now; you've managed teams and done a lot of fantastic work that I’ve followed through LinkedIn. There are often reports of high failure rates for AI projects. What do you think are some of the common challenges that organizations face that could contribute to these failures?
DT: I’ve managed quite a few teams at different organizations, and they were pretty successful from my perspective. I know people for whom it’s been tougher, though.
The most common pattern I’ve seen is companies making the wrong investment or doing the wrong thing at the start of their AI or data science project. They start by hiring a lot of people because that is what is common. They want to build a new capability, so they hire a lot of people. They allocate a very big budget to that.
But what happens is that you have a lot of new people with no industry experience. A lot...