Senior leaders set the AI ambition. Technology teams select platforms. Consultants and vendors can help build the infrastructure. Yet the moment AI reaches daily work, a different group takes control of whether the strategy lives or dies: middle managers.
This is especially important in the Gulf, where AI plans are moving quickly from national vision to operational deployment. The UAE’s federal Agentic AI framework requires each ministry and entity to form implementation teams that connect leadership, government services, operations, digital transformation, and AI. That cross-functional design is sensible. The harder challenge begins one layer down, where managers translate strategy into the routines, incentives, approvals, and conversations employees experience every day.
Managers Sit in the Adoption Squeeze
Middle managers face pressure from both directions. Executives want visible progress and measurable gains. Employees want to know what AI means for workload, job security, professional value, and accountability when something goes wrong. The manager is expected to deliver transformation while keeping the team productive during the transformation.
In focus groups with more than 250 middle managers conducted during research for my book, I repeatedly saw the same practical concerns. Managers worried about being held responsible for AI mistakes, lacked time to experiment, received ambiguous direction, and had to respond to employee anxiety while wondering what AI would do to their own roles.
Those concerns produce different behaviors. Some managers become AI Skeptics who demand overwhelming proof before changing a process. Others become Wait-and-See Traditionalists who prefer another department to go first. Enthusiastic Experimenters want to move but often lack authority or resources. AI Catalysts can accelerate adoption across units when leadership gives them room and support.
Give Managers Answers They Can Use With Their Teams
A strategy deck does not prepare a manager for Monday morning. Managers need concrete answers to questions employees will actually ask.
What outcome are we trying to improve? Which tasks are expected to change? Which decisions remain human? What information can go into approved AI tools? How will mistakes be reported and handled? What happens to saved time? How will employees get time to learn? What metrics will determine whether the experiment continues?
When managers cannot answer those questions, employees fill the gaps themselves. They may assume leadership is hiding a cost-cutting agenda, or that using AI creates personal risk with little upside. Clear manager guidance reduces that ambiguity.
Manager Support Is an AI Performance Variable
Microsoft’s 2026 Work Trend Index makes the management effect unusually visible. Its research found that organizational factors such as culture, manager support, and talent practices were much more strongly associated with reported AI impact than individual factors alone. Microsoft also found that many employees are moving faster than the organizations around them.
That mismatch turns the manager into either a bridge or a bottleneck. A supportive manager can give employees permission to test an approved tool, protect time for learning, help redesign a workflow, and normalize reporting failures. An unsupported manager can quietly block the same behavior by prioritizing short-term output, denying experimentation time, or treating every mistake as evidence that AI is unsafe.
Make Managers Owners of One Workflow at a Time
The most practical way to build manager capability is through responsibility for a real workflow. Give each manager one bounded process to improve with AI, a small team, clear guardrails, and a baseline metric. The manager should work with employees to identify the friction, test a new approach, decide where human judgment stays central, and report the outcome.
This approach is more useful than asking managers to become AI experts. Their comparative advantage is understanding the work, the people, the exceptions, and the consequences. They need enough technical fluency to supervise AI, but their central job is redesigning how the team operates.
Track Manager Readiness as a Leading Indicator
Organizations should also measure manager readiness directly. Do managers use the approved tools themselves? Can they explain the rules? Have they run a supervised experiment? Do their teams know where to escalate uncertainty? Are adoption and business results improving across their unit? Large differences between departments often point to management conditions rather than technology quality.
For Gulf organizations pursuing ambitious AI transformation, this is a useful warning. Executive sponsorship can launch an initiative, but it cannot substitute for management behavior at scale. The region’s AI strategy will become real inside thousands of team meetings, workflow reviews, performance conversations, and small decisions about whether an employee gets permission to try something new.
Middle managers are the transmission system. Strengthen that layer, and AI strategy reaches the work. Ignore it, and even strong technology and clear executive intent can stall between the boardroom and the frontline.
