Organizations have spent the last two years racing to embed artificial intelligence into decision systems, workflows, and day-to-day operations. In many places, that race has turned into pressure. Some employers now expect AI use as part of the job. A few have tied adoption to performance reviews. Workers in several industries say they worry that not using AI could affect their future at the company. The message is clear: AI is moving fast, and employees are told to keep up.
But the AI adoption track record leaves a lot to be desired. A 2025 EY Work Reimagined Survey finds that while almost nine in ten employees use AI at work, most keep it to basic tasks like search and summarization. A third worry that relying on these tools too much could weaken their own expertise. And even as workloads climb—something 64 percent say they’ve felt over the past year—only about five percent are using AI in ways that genuinely change how their work gets done.
The technology keeps advancing. Adoption hasn’t kept pace. And the slowdown has little to do with the algorithms themselves.
It has to do with managers.
New data from Gallup, my employer, shows how much manager support still shapes the AI experience. In organizations rolling out AI, only 28 percent of employees strongly agree their manager actively supports its use. The number reflects something deeper than comfort with a tool. Employees watch how their manager reacts before deciding what to do. If the response is tentative, the team slows down. If the manager seems unsure, employees read that uncertainty as a signal to wait. When the middle of the organization hesitates, AI stalls.
Executives can launch the initiative. Employees absorb it through their manager.
Start With the Managers, Not the Model
Employees who haven’t adopted AI explain a pattern leadership often overlooks. 44 percent don’t believe AI can help with their work. 11 percent prefer their current way of working. Ten percent have access but no understanding of how to begin. Eight percent have access but don’t feel safe using the tools at all.
These aren’t technical barriers. They’re interpretive ones. Employees want a manager who can explain the purpose of the tool, where it fits, and what responsible use looks like. Without that, the safest choice is to do nothing.
AI adoption isn’t about comfort with technology. It’s about clarity from the person they rely on to make sense of change.
Inside organizations that are putting real money behind AI, the difference a manager makes shows up sharply in the data. When employees strongly agree their manager supports the team’s use of AI, their behavior and belief shift. They are more than twice as likely to use AI regularly. They are over six times more likely to say the tools actually help them do their work. And the biggest shift sits underneath the surface: they are almost nine times more likely to feel that AI creates more room for them to use their strengths each day.
Support from a manager doesn’t just increase usage. It shapes how useful the tools feel and whether people see AI as something that expands their contribution rather than squeezes it.
Make AI Tangible in the Work People Already Do
AI becomes real when it ties into work employees already recognize. Many organizations announce AI with broad promises of transformation but leave employees guessing about what the change means for their actual tasks.
Managers play the bridging role here. Employees want to see how AI shortens documentation cycles, accelerates forecasting, reduces back-and-forth with customers, or removes repetitive steps that slow everything down. When managers can point to a specific workflow, bottleneck, or tedious step and show how AI shifts the load, teams begin to experiment.
Generalities rarely inspire action. Familiar moments do.
Remove the Safety Ambiguity That Silences Adoption
Access to AI doesn’t guarantee usage. Employees worry about what they might reveal or whether a dataset is allowed. Managers feel that pressure even more, since they carry responsibility for quality and compliance.
15 percent of employees cite legal, privacy, or compliance concerns as a reason they hold back. The hesitation often goes unspoken.
Managers need straightforward guidance. They need to know what data belongs in an AI tool, what must never be entered, when review is required, and how to handle uncertain output. Once these boundaries are clear, experimentation feels safe. And safety, not excitement, is often what unlocks movement.
Build Capability Around Judgment, Not Just Tools
Most AI training focuses on features. It walks people through menus and prompts. That approach misses what managers actually need.
Managers need the ability to read AI output with a critical eye. They need to understand what “good enough” looks like in their function. They need to know when to rely on the tool and when human review is essential. They need to weave AI into existing workflows without overwhelming the parts of the job that already work well.
This isn’t technical skill. It’s managerial judgment. And once managers build it, teams follow the tone they set.
Measure Progress Through Outcomes, Not Usage
AI adoption falters when organizations track the wrong indicators. Counting logins or prompt volume doesn’t show whether AI is improving anything. It just only shows whether people touched a tool.
Managers gain traction when the measures shift to things the team can feel: shorter turnaround times, fewer steps, less rework, improved accuracy, better customer responses. These outcomes give managers room to experiment, adjust, and improve without chasing usage metrics that don’t reflect real value.
Employees respond to changes that make the work better, not mandates.
Where Organizations Still Struggle
Even with clear data and practical steps, many companies face challenges that don’t disappear with guidance alone. Manager readiness varies far more than most AI strategies acknowledge. Some managers lean in quickly. Others feel exposed by the pace of change or lack the technical footing to lead their teams with confidence. Clarity helps, but capability often needs something deeper: hands-on coaching, time, and repeated practice.
Execution adds another layer. Leaders may agree that managers need support, yet scaling that support across distributed teams and varied roles is more complex than a single training program can solve. Turning manager enablement into something repeatable becomes a design issue, not a messaging one.
Context also shapes adoption. A manager in healthcare navigating HIPAA constraints works inside a different reality than a marketing manager experimenting with creative prompts. One environment is regulated. The other moves quickly by design. Universal guidance frames the work, but each setting demands its own calibration.
On top of that sits pressure from the top. Executives want movement, especially when competitors adopt AI at speed. That urgency lands directly on managers, who balance the push to “move faster” with the slower work required to build trust on the ground. When speed and readiness drift apart, change efforts splinter.
The urgency increases when AI use is framed as non-optional. Some employers now link AI expectations to performance reviews, and employees report feeling pushed into tools they don’t yet understand. When mandates outrun capability building, managers become the fuse. They carry executive pressure while also absorbing the responsibility to steady their teams. Instead of accelerating adoption, the mandate often creates quiet workarounds that hide capability gaps.
This tension shows up in the data. A 2024 Microsoft study found that while 79 percent of leaders say AI is essential for staying competitive, many are still unsure how to translate early gains into organizational impact. Fifty-nine percent say they struggle to quantify productivity improvements, and 60 percent report their company lacks a clear vision for how AI will be used. That uncertainty widens the gap between aspiration and execution. Employees, meanwhile, aren’t waiting. Seventy-eight percent of AI users are already bringing their own tools into the flow of work, shaping their day long before a formal strategy reaches them.
Different Managers, Different Starting Lines
Not every manager approaches AI from the same place. Some move toward it with curiosity, others with caution, and some with a blend of interest and self-protection. Treating all managers as a single group hides the very differences that shape adoption.
A small group jumps in early. They experiment on their own time, test ideas, and search for ways to take friction out of the team’s workflow. They don’t need motivation. They need guardrails and a place to share what they learn without overwhelming others.
Most managers sit in the middle—interested but hesitant. They see potential but worry about quality or because they aren’t sure how to explain missteps to their teams. Their hesitation isn’t resistance. It’s exposure. They move faster when the examples feel practical, when the learning curve isn’t steep, and when trying something small doesn’t feel like a gamble.
A smaller set feels resistant. Not hostile—just stretched. Some are overloaded. Others have built their identity around expertise they fear might lose value. These managers need space and pacing, not pressure. When the conversation shifts away from tools and toward the human parts of their job—judgment, coaching, sequencing, prioritization—they begin to re-engage.
Recognizing these differences doesn’t divide the management population. It lets organizations meet managers where they are, rather than asking everyone to start from the same line.
When Managers Gain Confidence, Teams Gain Momentum
A consistent thread runs through the data. Employees hold back when managers lack clarity. Managers hesitate when guidance is thin. AI struggles when both levels try to navigate uncertainty without a shared foundation.
Managers translate strategy into action. They set the tone for risk, pace, exploration, and caution. When they understand where AI fits, what it solves, and how to introduce it responsibly, teams begin to move. When that understanding is missing, adoption stays surface-level.
AI doesn’t reduce the importance of managers. It amplifies it. The technology depends on leaders who can help teams understand what the tool is for and how to use it well.
If organizations want AI to create meaningful value, the first investment isn’t in models or platforms. It’s in the people who make work legible to others.
The readiness of managers becomes the readiness of the organization. That’s where AI begins to succeed.