How Top AI Users Achieve Better Results

In the current business environment, artificial intelligence is no longer a novelty but a fundamental tool for competitive advantage. The critical challenge for organizations is not merely encouraging adoption, but fostering truly effective and impactful use. Recent research reveals a significant gap: while most employees now interact with AI, only a small fraction leverage it to its full potential. The habits of these top performers, however, provide a clear blueprint for success that others can follow.

A study conducted by KPMG and the University of Texas at Austin’s McCombs School of Business analyzed over 1.4 million real-world AI interactions across the professional services firm. The data showed that 90% of the 2,500 employees studied used AI, but a mere 5% were classified as highly sophisticated users. These individuals didn’t just use AI more often, they used it differently. Their approach is defined by specific, teachable behaviors that drive superior outcomes.

First, they engage in longer, more interactive sessions. They treat the technology as a dynamic partner, engaging in extended dialogue with multiple rounds of feedback and refinement. This stands in contrast to users who input a single prompt and accept the first output. Second, they push back and iterate. They don’t settle for initial results, instead providing clear roles, examples of desired outputs, and detailed constraints to guide the model toward a better answer. Third, they embrace complexity. They confidently delegate multi-step, intricate tasks to AI, providing the structured guidance needed for sophisticated work. Finally, they use AI for brainstorming and exploration, not just for automating simple tasks. Notably, their interactions often feature a conversational, informal tone, signaling a more fluid and collaborative relationship with the tool.

Zach Kowaleski, an assistant professor at UT and co-author of the research, encapsulates the mindset of these sophisticated users. “Frequency, repetition helps. Ambition, ask for more. Persistence, don’t settle for the first response. Flexibility, play with different models so you get familiar with the different advantages they offer.”

Simply providing access to AI tools is insufficient to cultivate these high-impact behaviors. The study concluded that meaningful, value-creating use does not spread organically. Many leaders mistakenly approach AI implementation as a standard software rollout, failing to recognize it as a fundamental shift in how work is conceptualized and executed. This perspective is echoed by media theorist Douglas Rushkoff, who compares the societal impact of contemporary AI to the invention of writing itself. He argues that the most valuable application of this technology is not for finding definitive answers, but for asking better questions and engaging in a generative, exploratory practice.

To bridge the skills gap, intentional training and cultural design are essential. A World Economic Forum report highlights that nearly 60% of workers will require reskilling by 2030 to keep pace with technological change. The researchers from UT and KPMG recommend several actionable strategies for organizations. They advise codifying AI-first best practices through internal playbooks and peer networks that demonstrate sophisticated use in context. Investing in hands-on, scenario-based training built around real client work and internal tasks helps build practical confidence. Furthermore, setting role-specific expectations clarifies what “good” AI-assisted work looks like for different positions, giving employees a tangible target beyond basic usage.

The ultimate goal is to systematically transform ordinary users into high-impact ones. This requires moving beyond celebrating the top 5% to actively designing the support systems that enable everyone to elevate their practice. Organizations that succeed in this cultural and operational shift will not only enhance productivity but will secure a decisive edge in an AI-driven economy.