A company can give employees access to ever more capable AI and still fail to achieve a corresponding business benefit. Technology also accelerates poorly designed processes, unclear authority and unnecessary coordination. Managers then receive more materials and requests without gaining more time to interpret them. The real problem therefore lies not only in the speed of AI adoption but in the slowness of organizational change.
Discussion of AI adoption often focuses on technological speed: which models are newer, what they can automate and how many employees already use them. Keith Ferrazzi points to a different problem. Technology is changing faster than the way companies are organized.
AI therefore acts as a stress test. If an organization has not clearly allocated decision rights, the system will produce more material but will not accelerate decisions. If a workflow contains unnecessary approvals, AI may speed up preparation of every step while the overall process remains slow. And if teams do not share rules for setting priorities, a larger volume of generated information may mainly create more work.
Ferrazzi links this phenomenon to an “AI confusion tax”: technology increases production capacity while integration and interpretation work remains with people. This pressure may be felt most strongly by managers, who receive ever more proposals, analyses and requests and must decide which of them are actually important.
The problem, therefore, is not that AI fails to save time in individual activities. The saving may simply move elsewhere. An analyst prepares material faster, but the manager receives twice as many analyses. A salesperson creates more personalized proposals, but their manager has to review a larger number of activities. A team can generate project variants faster, but the decision committee then has to assess more of them.
When evaluating the benefit of AI, it is therefore not appropriate to measure only the time required to perform an individual task. Follow the entire flow of work. Has the time from the emergence of a problem to a decision shortened? Are there fewer handoffs? Has the number of approval steps fallen? Can people make better decisions from the larger amount of information, or merely make decisions more often?
The second step is to remove work that no longer makes sense. If AI can prepare ten reports instead of two, that does not mean leadership needs ten reports. A more useful question is whether some of them can disappear entirely.
Managers therefore need the ability to restructure work alongside the ability to use AI. This is not only about training on new tools. Leaders must be able to eliminate activities, simplify decision authority and prevent higher information output from overwhelming the rest of the organization.
KEY TERMS
- AI confusion tax: Additional coordination and interpretation created when AI is introduced rapidly into an organization that has not changed.
- Organizational readiness: A company’s ability to adjust processes, authority and management so that technology genuinely creates value.
- Integration work: The work required to combine a larger volume of information and outputs into a usable decision.
- Process stress test: A situation in which new technology exposes existing weaknesses in a work process.