The old model of a human simply approving an AI's output is obsolete. A "human at the helm" approach is needed, where leaders strategically decide where and when to insert human judgment into complex, multi-step AI workflows, acting as a director rather than a final gatekeeper.
When designing AI customer service, consider the customer's emotional state, not just task complexity. Evidence suggests angry customers want human interaction, while embarrassed ones may prefer AI's anonymity. This emotional lens in system design prevents poor business outcomes and improves customer experience.
To prevent "cognitive offloading" where employees blindly trust AI, design systems with deliberate friction. This forces a pause, encouraging users to apply their own judgment and take full accountability for the final output. It is crucial for high-stakes decisions where AI can make mistakes.
AI tends to revert to the mean and produce generic ideas. For true breakthroughs, innovation teams must first brainstorm unique concepts without AI to define a novel direction ("front-load the brief"). Only then should they use AI as a powerful co-ideator to expand on those creative sparks.
As AI automates traditional entry-level tasks, the role of new graduates is shifting. Instead of doing routine work to learn the business, their value now lies in mastering AI tools and managing AI-driven processes, often teaching senior colleagues more effective ways to leverage the technology.
Since no one has decades of experience in emerging AI skills, traditional hiring metrics like company pedigree are failing. Leaders need a "bias toward the future," evaluating candidates on their demonstrated ability to create and solve new problems rather than relying on outdated resume shortcuts.