AI adoption is accelerating inside large organizations, but employee guidance is not keeping pace. New research from KPMG shows that companies are moving from AI experimentation toward broader workplace use while still developing clear rules for how employees should work with the technology.
KPMG's Q3 2026 AI Quarterly Pulse Survey, released in September, surveyed 314 U.S. C-suite and business leaders at companies with annual revenue of at least $1 billion.
The survey found that 44% of organizations now report significant workforce adoption of AI, up sharply from 10% a year earlier. At the same time, 62% are building, deploying or developing AI agents, showing how quickly AI is moving into operational work.
But the policies surrounding that adoption are developing more slowly.
AI adoption is moving from pilots to everyday work
KPMG found that 74% of organizations now include cost reviews in their AI approval processes, up from 61% in the previous quarter.
Other forms of oversight are also becoming more common. About 70% of organizations use AI monitoring dashboards, while 43% have introduced usage or token budgets to control how much AI processing teams can use.
The findings suggest that organizations are becoming more deliberate about managing AI as deployments expand.
That shift is particularly important as companies move beyond isolated pilot projects. KPMG AI enterprise transformation leader Rahsaan Shears said the financial impact of AI can become more significant when a system moves from experimentation to large-scale use.
At the same time, governance is becoming more established. Seventy-three percent of leaders said they were confident in their organization's ability to govern AI, compared with 57% in the previous quarter.
Nearly half, 49%, said their organizations have defined high-risk use cases where autonomous decision-making by AI agents is not permitted.
Employee guidance is still lagging
While organizational governance is becoming more mature, employees may not have the same level of clarity about how AI should be used in their day-to-day work.
KPMG found that 55% of organizations are developing role-specific guidelines for working with AI agents, up from 21% a year earlier.
The increase is significant, but it also means that nearly half of the organizations surveyed have not yet reached that stage.
Employee research points to an even wider communication gap.
According to Gallup's May 2026 tracking of workplace AI use, only 25% of U.S. employees said their organization had communicated a clear plan for integrating AI.
A separate Traliant report covered by HRD America in July found that 51% of HR teams reported having clear AI-use guidelines and employee training.
The difference between executive-level confidence and employee-level clarity could become increasingly important as AI becomes part of regular workflows.
Employees need to know not only that AI is available, but also where its use is appropriate, what information can be entered into AI systems, when human judgment is required and how AI-assisted work will be evaluated.
The role of HR is changing
As AI becomes more embedded in jobs, governance is no longer only a technology issue.
Who sets workplace AI rules varies between organizations, according to Shears. Depending on the company, the CEO, CFO, CHRO or technology leadership may take the lead.
For HR teams, the issue extends into workforce planning, job design, performance expectations and employee communication.
The challenge is also becoming less about introducing AI and more about defining how people and AI will work together.
Organizations that provide employees with clear expectations may have an easier path to broader adoption than those that make AI widely available without explaining how it should be used.
The bigger shift
The latest KPMG findings show that AI is becoming increasingly operational for large employers. Adoption, monitoring and governance are all advancing, but employee guidance is still developing.
For HR leaders, the next stage of AI adoption may therefore involve more than choosing tools or measuring productivity. It will also require clearer expectations around how employees use AI and how work changes as a result.
The technology may be moving quickly, but organizations still need to determine how people should work with it.
