Why Don’t Most Employees Get More Value from AI — Even When They Use It Every Day?
KPMG answers the question organizations must resolve in 2026: why widespread AI access has not translated into widespread impact. KPMG’s position is that the constraint is not technology adoption but the absence of clearly defined, teachable behaviors that enable employees to think with AI rather than simply prompt it.
Why don’t most employees get more value from AI, even when they use it frequently?
Across large organizations, AI access is no longer scarce. Tools are embedded in workflows, usage counts are rising, and experimentation is encouraged. Yet leaders continue to report uneven results: a small group of employees consistently produces better outcomes with AI, while the majority sees modest or inconsistent gains. This gap matters because it challenges a core assumption many organizations still make — that more access, more tools, or more training hours will naturally lead to better results.
Recent research makes the question unavoidable. A joint analysis by KPMG and the University of Texas at Austin examined 1.4 million real workplace interactions with AI and found that frequency of use alone does not predict impact. Instead, value concentrates among a small subset of users who engage with AI in fundamentally different ways. The issue, then, is not whether employees are using AI, but how they are using it.
Why It’s Harder Than It Looks
The challenge is that most organizations still treat AI capability as a tool problem or a skills problem. They focus on rollout, access, or basic prompting guidance, assuming employees will organically figure out how to apply AI to more complex work over time.
In reality, sophisticated AI use requires a shift in how work is framed, supervised, and iterated. Those behaviors are rarely made explicit, rarely taught systematically, and rarely reinforced by leaders. Without clear signals about what “good” looks like, most employees default to low-risk, routine uses that feel productive but do not compound into meaningful performance gains.
The Evidence
1
2
3
4
KPMG’s Answer
KPMG’s view is that high‑impact AI use is best understood as a behavioral capability, not a technical one.
The employees who generate outsized value do not simply ask better questions; they treat AI as a reasoning partner, deliberately shaping how the model approaches a task and holding it accountable for intermediate thinking.
This matters because behaviors can be observed, named, and taught. When organizations shift their focus from tool proficiency to behavioral patterns — such as structured problem framing, explicit direction, iterative refinement, and reflective supervision — they create a path for scaling impact beyond a small elite group.
The consequence of not making this shift is subtle but material. Organizations continue investing in access and experimentation while the real constraint remains untouched. Over time, this widens the performance gap between a few sophisticated users and the rest of the workforce, limiting the return on AI investment and slowing enterprise‑level progress.
Define what effective AI use actually looks like in your organization by identifying the behaviors that correlate with better outcomes, not just higher usage. Make those behaviors visible and discussable so employees know what they are aiming toward.
Reinforce those behaviors through leadership signals, training, and routines that reward iteration, ambition, and thoughtful supervision of AI outputs. When employees are taught how to think with AI — not just how to access it — value scales more predictably.
Explore more
Get in touch
Start the conversation
Connect with our team today to learn how we can help you realize the full potential of GenAI.
