Why are so many organizations investing millions into AI while still struggling to prove meaningful productivity gains?

In this episode of AI at Work, I spoke with Rahsaan Shears, Principal and AIQ Program Lead at KPMG, about a major new study conducted alongside the McCombs School of Business at The University of Texas at Austin that analyzed 1.4 million real workplace AI interactions. What emerged from that research challenges many assumptions business leaders currently hold about AI adoption, productivity, and the future of work.

One of the most surprising findings was that the most effective AI users were not necessarily the most technical employees, nor even the people using AI tools most frequently. Instead, the highest performers were what KPMG calls “sophisticated users,” employees who learned how to think with AI, challenge it, iterate with it, and use it as a reasoning partner rather than simply a faster search engine.

Rahsaan explained how this distinction is forcing organizations to rethink how they measure AI success. Many businesses remain focused on surface-level adoption metrics like license counts, prompt volume, or chatbot usage. But those measurements often fail to capture whether AI is genuinely improving decision-making, productivity, creativity, or operational performance. The real challenge, according to Rahsaan, is that most organizations still lack a framework for understanding what meaningful AI-enabled work actually looks like.

We also explored the growing behavioral capability gap emerging inside organizations. While some employees are rapidly learning how to integrate AI into their workflows in sophisticated ways, others remain stuck using these tools for basic task acceleration. Rahsaan shared why this gap has less to do with age or technical skill and far more to do with curiosity, ambition, critical thinking, and an employee’s willingness to rethink how work itself gets done.

One of the strongest themes throughout our conversation was the idea that AI should not be treated as a technology rollout alone. Rahsaan argued that organizations succeeding with AI are redesigning culture, workflows, decision-making structures, and team dynamics at the same time they deploy new tools. He compared today’s AI systems to toddlers: incredibly capable compared to where they started, but still requiring guardrails, coaching, supervision, and careful integration into everyday work.

For listeners interested in organizational transformation, this episode offers practical insight into how KPMG is building AI-first behaviors through peer-led champion networks, embedded learning models, AI coaching inside the flow of work, and safe environments where employees can experiment without fear of failure. Rahsaan shared why psychological safety, curiosity, and continuous learning are rapidly becoming core business skills in the AI economy.

We also discussed why organizations that fail to create agency for employees may struggle to scale AI beyond pilot programs. According to Rahsaan, many existing business processes were designed around the limitations of human workers, limitations that no longer fully apply once digital teammates and agentic workflows enter the picture. Companies willing to question long-standing assumptions about work itself are beginning to separate themselves from the rest of the market.

This conversation moves beyond AI hype and focuses on the human behaviors, organizational structures, and operational changes that will ultimately determine who wins and loses in the AI economy.