The Role of AI in the Future of Work: A Conversation with Jeremy Utley | APQC 2025 Conference

AI’s biggest workplace impact isn’t that it replaces people—it accelerates whatever cognitive habits already exist, then reshapes how knowledge and process work get done. Jeremy Utley argues that technology acts like an amplifier: when teams understand common thinking traps, AI helps them bypass them faster; when they don’t, AI can speed up mediocre decisions by locking people onto early, low-quality ideas. A key example is the “Einsteining bias,” a tendency to fixate on an initial solution regardless of quality. With AI, the path to both good and bad ideas can become faster—so the real differentiator becomes metacognitive awareness: knowing that humans gravitate toward quick, easy answers and using AI in ways that counter that pull (for instance, prompting for many ideas while actively resisting premature fixation).

That bias problem connects directly to how organizations should integrate AI into everyday collaboration. Utley describes “knowledge cafe” style sessions—human-led idea-building with rotating participation—then suggests feeding the outputs into AI for expansion, devil’s-advocate critique, or role-based questioning (including asking for missing expert perspectives). The goal isn’t just more brainstorming; it’s creating a feedback loop where communities of practice share use cases and spark new ones. In that environment, AI becomes a tool for widening the range of possibilities rather than narrowing attention to the first acceptable answer.

Utley then ties AI to knowledge management and process management, arguing that generative AI can function as an expanded short-term memory for organizations. Instead of relying on what’s readily accessible in a person’s head or a static wiki, teams can query richer context—helpful when searching for examples, contract language, or prior decisions. He illustrates this with his own experience reading Idea Flow: he had to pause and take notes because he couldn’t reliably recall stories and research later. He also cites Guy Kawasaki’s “Kawasaki GPT,” built from Kawasaki’s books, which can answer questions about his work even when Kawasaki himself doesn’t immediately know the response.

The practical payoff, Utley says, shows up in cycle-time reductions when process knowledge is paired with AI assistance. He offers two examples: a fashion design workflow that reportedly shrank from two months to two minutes for concept-to-sample-to-photo-shoot; and work with the National Park Service, where facilities managers preparing documentation for statements of work and proposals reportedly cut turnaround time from two or three days to about three hours, with one tool prototype built in 45 minutes. Utley estimates the broader impact could save thousands of labor days, and he emphasizes that these gains don’t require technical expertise—just familiarity with the process and basic generative AI skills.

Finally, Utley reframes influence and change management for the AI era. Every individual can access an assistant, expert, coach, and creative partner, so leadership becomes less about authority and more about enabling people to use AI effectively. He warns that “silence” from leaders increases fear and uncertainty, while the real job is clarifying expectations and redefining AI as “augmented I”—work guided and steered by humans with AI as support. His actionable advice is simple: experiment daily (even 15 minutes), ask “Have you tried ChatGPT?” often, and treat use-case discovery as an ongoing practice rather than a one-time rollout.