AI at work has reached a tipping point. New research shows broad adoption, but a lack of clarity is holding employees back.
Amanda Wowk
Senior Research Writer
Workday
TL;DR: As AI shifts from answering prompts to proactively doing work behind the scenes, employees need help understanding what agents can actually do.
In a September 2026 survey of nearly 6,000 workers across the globe, 69% of employees using AI say they’ve allowed AI to complete a whole task at least once. However, senior leaders are 2.5x more likely than individual contributors to say they understand AI agents well or could confidently explain the term.
Many employees lack an in-depth understanding of what AI agents are—nevermind what they can do. Just 38% of all employees say they understand the term “AI agent” well enough to explain it confidently.
Senior leaders, on the other hand, are much more likely to understand the term, and to have delegated a whole task to AI. Nearly nine in 10 (88%) have used agentic AI to complete tasks versus just 54% of individual contributors—a difference of 34 points.
As companies move from AI assistance toward agentic AI that works proactively behind the scenes, these gaps will matter. Leaders may be comfortable with total task delegation, but many employees are still figuring out what AI agents are, when to use them, and how to use them responsibly in their daily work.
So, what’s holding employees back from going deeper with AI? Let’s dig further into the findings:
Report
Employees are now seeing AI as a source of value, not just change. The share of employees who say AI makes their day-to-day work better (versus worse) rose seven points month over month—the largest positive shift across Workday’s AI@Work metrics. But the benefits of AI are not being experienced equally within organizations.
Senior leaders report far stronger adoption than individual contributors, according to Workday’s September survey data. Six in 10 individual contributors (62%) report AI adoption, compared with 95% of senior and executive leaders. AI power users are most prevalent at the top of companies, too: 44% of senior leaders report heavy AI use, compared to 9% of individual contributors.
At the same time, leaders are about twice as likely as individual contributors to trust AI and say it is improving their work. Individual contributors report a +30 work-experience score, which measures the percent who said AI made day-to-day work better minus the portion saying it made it worse. This is compared with a score of +65 among senior and executive leaders. Only 32% of individual contributors trust their employer’s approach to AI, versus 66% of senior and executive leaders.
These gaps suggest that senior leaders may be experiencing AI as a strategic advantage, while individual contributors are still navigating more practical questions around access, trust, quality, and how AI fits into daily workflows.
The shift: If AI is to become a true workforce advantage, organizations must ensure that the people closest to day-to-day operations have the same clarity, support, tools, and permission to experiment as senior leaders do.
As AI improves the work experience, employees are still not fully comfortable disclosing their AI use. In fact, 41% of employees who use AI say they deliberately conceal their use. Last month, 49% of employees said they had concealed AI use.
The data shows employees still need a clearer understanding of AI policies. Only 35% of employees say their employer’s expectations for disclosing AI use are very or extremely clear. Nearly one in five (19%) say their employer has no disclosure expectations at all.
This is the concealment conundrum: employees are using AI broadly, but many still lack a shared understanding of the rules so they are afraid to speak up.
The shift: Replace abstract AI policies with practical, role-based guidance. Employees need concrete examples of approved uses, disclosure expectations, data boundaries, and human-review requirements.
Leaders can look to Germany for a lesson in transparency. The data shows concealment fell sharply, by 23 points, to 39%—the largest country-level drop. Germany’s more explicit norms, dialogue, and local workplace conditions are likely making AI use easier to acknowledge openly.
AI can save employees time, but those gains are often offset by the work required to review and fix its output. Across the globe, heavy AI use held steady at about 21%. But one in three employees (33%) continue to say they lose at least half of the time AI saves them to rework.
Trust in AI output has also stalled month over month. Only 44% of employees say they can rely on AI outputs without closely checking them—a sign that many still see AI as a helpful starting point rather than a dependable collaborator.
This creates a practical barrier to deeper use. Employees are unlikely to hand more complex or higher-stakes tasks to AI if they expect to spend much of the saved time reviewing, correcting, or rebuilding the work themselves. The issue is especially important as organizations encourage more agentic AI use, where AI takes on a greater share of a task from start to finish.
The shift: Organizations must help employees use AI where it adds the most value, rather than asking them to apply it to every task. That means identifying work that is repeatable, lower risk, and easy to review; giving employees clear examples of when AI is appropriate; and setting expectations for human oversight before AI-supported work is shared or acted on.
Here’s what else we’re tracking month over month. Come back in November for new insights and trends.
| KPI | What it measures | August results | September results | Month-over-month change |
|---|---|---|---|---|
| AI Adoption Rate | % of employees using AI for any part of their work, excluding “none of it” and “I don’t use AI at work” | 76% | 77% | up +1 pt |
| Power-User / Heavy-Use Rate | % using AI for “most” or “almost all” of their work | 21% | 21% | +0 pts |
| Agentic Task Delegation Rate | % of AI users who have let an AI tool complete a task fully autonomously with little or no input at least once | 71% | 69% | down -2 pts |
| Average Weekly Time Savings | Mean hours per week employees say AI saves them, using bucket midpoints | ~3 hours/week | ~3 hours/week | No change |
| AI “Rework Tax” | % of AI users who say 50%+ of their time savings is eaten up fixing or reworking low-quality AI output | 33% | 33% | No change |
| Trust in AI Output Accuracy | % who agree or strongly agree they can rely on AI outputs without close checking | 43% | 44% | up +1 pt |
| Shadow AI / Concealment Rate | % of AI users who deliberately hid their AI use on a work task | 49% | 41% | down -8 pts |
| AI + Work Experience Sentiment | Net score: % saying AI made day-to-day work better minus % saying it made it worse | +36 pts | +43 pts | up +7 pts |
| Employer AI-Trust Score | % who somewhat or strongly trust their employer to use AI in a way that benefits employees, not just the organization | 42% | 45% | up +3 pts |
| Career Confidence Gap | Net career optimism: % agreeing AI will create new opportunities minus % agreeing AI will make current skills less valuable | +4 pts | +6 pts | up +2 pts |
| Training | % of employees who have received AI training from their employer | 81% | 80% | down -1 pt |
| Manager AI Conversation Rate | Share with a manager who discussed AI role impact in past month | n/a | 56% | n/a |
Report
A gap between effort and impact is the real ROI problem. Discover the solution.