Here’s a bold statement: AI automation isn’t the plug-and-play solution many leaders think it is. Despite the hype, even the CEOs of multi-billion-dollar AI companies are sounding the alarm. But here’s where it gets controversial—while AI promises to revolutionize workflows, the reality is far messier than anyone expected. Let’s dive in.
In a recent episode of the Bg2 Pod, Glean CEO Arvind Jain and Databricks CEO Ali Ghodsi shared eye-opening insights about the challenges of implementing AI. Jain, whose company raised $150 million at a $7.2 billion valuation, admitted that even simple-sounding tasks—like using AI to identify employees’ weekly priorities—proved far more complex than anticipated. “I thought it would magically work,” he said. Spoiler alert: it didn’t. Ghodsi echoed this sentiment, stating, “You can’t just unleash AI and expect it to work seamlessly.”
And this is the part most people miss: AI automation requires meticulous engineering, strong teams, and a lot of patience. Jain highlighted a failed attempt to build a custom AI model for a specific use case, which ultimately led them back to relying on existing foundation models. “Success takes much longer than you’d think,” he noted. Ghodsi’s Databricks, valued at $134 billion after a $4 billion funding round, faces similar hurdles. Both CEOs agree that failed AI projects aren’t failures—they’re signs of experimentation. “If 95% of your projects fail, you’re on the right track,” Jain explained. “It means you’re pushing boundaries.”
But here’s the real kicker: humans aren’t going anywhere. Ghodsi emphasized that human oversight will remain essential, even as AI takes on more tasks. “We’ll all become supervisors,” he predicted. This idea is backed by AI pioneer Yoshua Bengio, who argues that uniquely human qualities—creativity, empathy, and intuition—will only grow in value as automation advances. “The human touch will become more precious,” Bengio said on The Diary of a CEO podcast.
So, here’s a thought-provoking question for you: Is AI automation overhyped, or are we simply underestimating the effort required to make it work? Share your thoughts in the comments—let’s spark a debate!
