Companies have rapidly rolled out AI tools, but many leaders still struggle to measure whether they’re actually improving performance, Harvard Business Review writes.
New research based on 1.4 million AI interactions suggests the problem isn’t adoption, but how employees use the technology.
The most effective users treat AI as a collaborative thinking partner, not a shortcut. They write structured prompts, iterate on responses, switch between tools intentionally and delegate complex, multi-step tasks with clear objectives. In contrast, many organizations still track surface-level metrics like usage frequency, which reveal little about real impact.
As Harvard Business Review notes, the gap is significant: While AI adoption can exceed 90%, only a small share of employees demonstrate truly sophisticated use.
The takeaway for leaders is clear. Driving value from AI requires shifting focus from access to behavior, defining what “good” looks like and training employees to think more critically, strategically and creatively alongside these tools.
Read the full story from the Harvard Business Review. A subscription may be required.
