Here’s a startling fact that should make every business leader pause: companies are funneling a staggering 93% of their AI budgets into technology and a mere 7% into the people who are supposed to use it. According to Bill Briggs, Deloitte’s Chief Technology Officer, this lopsided investment strategy is not just surprising—it’s a critical error that could derail the potential of AI to transform industries. But here’s where it gets controversial: while the world worries about robots stealing jobs, the real issue might be how companies are neglecting the human element in their tech-driven strategies.
In a recent conversation with Fortune at Deloitte’s New York City office, Briggs shed light on this alarming trend. He likened the current approach to “buying all the ingredients for a gourmet meal but forgetting to follow the recipe.” Companies are obsessing over AI models, chips, and software—the flashy components—while overlooking the culture, training, and workflows needed to make the technology work. As Briggs puts it, “It’s like trying to make paella and ending up with just cilantro.”
This imbalance isn’t just a theoretical concern; it’s rooted in real-world observations. Briggs, who has been at the forefront of Deloitte’s Tech Trends report for nearly two decades, admits he felt this disparity during his travels but hadn’t been able to quantify it until now. He calls it a trap of “incrementalism”—the tendency to slap new technology onto old processes instead of reimagining how work gets done. And this is the part most people miss: AI isn’t just another tool to bolt onto existing workflows; it demands a holistic transformation of how organizations operate.
Briggs doesn’t mince words when he says the most damaging phrase in any organization is, “We’ve always done it this way.” This mindset, borrowed from computer science pioneer Grace Hopper, is exactly what’s holding companies back. To truly harness AI’s potential, leaders must push beyond their comfort zones and rethink everything from job roles to performance management.
But here’s the kicker: the consequences of ignoring this human-centric approach are already showing up in the workforce. Deloitte’s TrustID report reveals that despite increased access to generative AI (GenAI), usage has dropped by 15%. Even more alarming is the rise of “Shadow AI”—a phenomenon where 43% of workers admit to using unapproved AI tools, often because they find them easier and more accurate than company-sanctioned options. This isn’t just a compliance issue; it’s a trust issue. Corporate worker confidence in GenAI plummeted by 38% between May and July 2025, while those who received hands-on AI training reported 144% higher trust in their employer’s AI systems.
So, what’s driving this reluctance to invest in people? Briggs points to a deep-seated fear of “buyer’s remorse.” CEOs and boards are terrified of committing to a vendor or technology only to see a better option emerge days later. It’s like trying to time the stock market perfectly—a strategy that rarely pays off. Briggs compares this hesitation to a “pre-snap penalty” in sports: the longer you wait to act, the further behind you fall.
The urgency to correct this 93-7 imbalance is only intensifying with the rise of “Physical AI,” which extends beyond text generation to robotics and drones. Real-world success stories, like HPE’s 50% faster reporting after deploying Zora AI, prove that getting the human-tech integration right can yield massive returns.
But here’s the bold question: Are companies willing to rethink their entire approach to AI, or will they continue to prioritize shiny new tools over the people who use them? Briggs’ message to the C-suite is clear: the technology is ready, but without a focus on cultural and human transformation, even the most advanced AI will gather dust. As he aptly warns, “No matter how much traffic there is, the sooner you leave, the sooner you can get there.”
So, what do you think? Is the 93-7 split a necessary evil, or a recipe for disaster? Let’s debate this in the comments—because the future of work depends on getting this right.
