Two top AI executives, Databricks CEO Ali Ghodsi and Glean CEO Arvind Jain, have warned that automating work with AI is far more complex and challenging than many business leaders assume. Speaking on the "Bg2 Pod" episode released Tuesday, the two CEOs shared candid insights about the realities of deploying AI in real-world organizations. Jain, who leads Glean—a company that uses AI to help employees search across internal tools and documents—recalled an ambitious internal project to use AI to automatically identify each employee’s top weekly priorities and compile them for leadership. “It has all the context inside the company to make it happen,” he said, adding that he initially believed the task would be simple, even “magical.” In practice, it didn’t work. The effort failed to deliver reliable results, underscoring the gap between theoretical potential and practical execution. Glean recently raised $150 million at a $7.2 billion valuation, but Jain admitted that even high-stakes internal AI experiments can fall short. He also described a failed attempt to build a custom AI model for a specific product function. That project ultimately didn’t succeed, forcing the company to pivot back to using existing foundation models, which, while less tailored, were more reliable and easier to implement. “It actually takes much longer than you know to actually generate success,” Jain said. Ghodsi, whose company Databricks provides a data and AI platform used by major enterprises, echoed the sentiment. “It's not just you can just unleash the agents, and it just works,” he said. He emphasized that making AI effective within an organization is not a plug-and-play process, but rather a complex engineering challenge that requires deep integration, careful testing, and strong teams to manage and maintain. Databricks recently announced a $4 billion funding round, valuing the company at $134 billion, highlighting the massive scale of investment in AI infrastructure. Yet, both leaders stressed that failure is not a sign of poor strategy—it’s often a natural part of innovation. “You hear these 95% of projects fail,” Jain said. “That's actually what you want. When you're actually experimenting with new technology, if all of your projects are failing, that means you're not trying enough.” Both executives also stressed the enduring need for human involvement. Ghodsi has previously argued that even as AI agents take on more tasks, humans will remain essential as overseers. “In a few years, yes, we'll have agents in many, many places, but there will be a human overseeing and approving every step,” he said. “You're on the hook when you approve, when you click, 'OK.' We all become supervisors.” This view is shared by other leading figures in the field. Yoshua Bengio, a pioneer in AI and one of the field’s most respected researchers, has said that as machines take over routine tasks, human qualities like creativity, empathy, and judgment will become increasingly valuable. “Work on the beautiful human being that you can become,” he said on a recent episode of “The Diary of a CEO” podcast. “I think that part of ourselves will persist even if machines can do most of the jobs.” He added that the human touch will gain greater importance as automation advances.