Caught in the AI Crossfire: 5 Moves for Middle Managers to Start Shaping the Future
McKinsey’s new AI report reveals a gap between ambition and reality. Here’s what it means for the managers in the middle, and why they’re critical to getting it right.
By Elina ·
A Tension in the AI Conversation
I recently readSuperagency in the Workplace, a 2025 McKinsey report on how both employees and C-suite leaders see AI reshaping the workplace.
The data paints a familiar but uneasy picture: there’s enormous appetite for AI adoption, but little clarity on who’s leading the charge or where exactly we’re headed. The pace of change is fast, but the roadmap is murky.
For mid-level managers, this creates a double bind. You’re expected to keep operations stable while also preparing for a technological disruption. You’re being told AI will transform your work without much guidance on how, when, or what skills will keep you indispensable.
So I read the report through a lens: what does all this mean for the leaders in the middle, those neither setting the company-wide C-suite level strategy nor simply carrying it out, but tasked with bridging the two?
My Takeaways from the McKinsey Data
Here’s what stood out to me, and why it matters for mid-level leaders.
1. Appetite Is High, But the Path Is Unclear
92% of organizations plan to increase AI investments, and employees are signalling readiness, as nearly half say they want formal training and access to AI tools. Yet only 1% of companies have AI fully integrated and delivering measurable results. This gap between eagerness and execution, shows that we are in a transitional zone.
In transitions like this, benchmarks and best practices offer little guidance, because they don’t exist yet. For mid-level leaders accustomed to steering by precedent, this can feel disorienting. But it’s also an opportunity.
In times of ambiguity, your role is less about having the answer, and more about defining principles that help you act without full certainty. The question becomes: What is my relationship with not knowing? How do I move forward when the map is still being drawn?
2. ROI Is Elusive
While investment is booming, measurable returns are rare. Only 23% of companies report cost reductions from AI investment, and just 12% have identified scalable, value-generating use cases.
The technical challenges matter, but the deeper issue is that AI is often pursued as a goal in itself, rather than a means to serve the business. To integrate AI meaningfully, leaders need to know their business (and themselves) so well that they can take it apart, reimagine it, and rebuild it with the new tools.
The barriers aren’t just technical-legal, compliance, and ethical risks are still to be mitigated with deep understanding of nuances, employees, customers and systems.
Another key challenge is adaptability: Can companies pivot quickly enough as the tech evolves? That will depend on the leadership capacity throughout the organization, not on just brilliance of external consultants, who defined the roadmap or technical tools.
Strangely, becoming skilled in effective AI roll-out will likely to require becoming more of an expert in human systems (customers, teams, processes) and ability to stay connected with the core value delivery, instead of getting distracted by the buzz words and novelty.
3. The “Who” Problem: Leadership Paralysis
Mid-level leaders, especially millennials aged 35–44, are emerging as the most enthusiastic and experienced AI users: 62% report high usage, compared to 50% of Gen Z and just 22% of baby boomers.
Yet leadership perceptions lag. Executives estimate only 4% of employees are using AI for work regularly, when employees themselves report 12%.
This mismatch fuels hesitation at the top, and uncertainty below. Leaders delay big moves, employees wait for direction, and pilots stall.
In that vacuum, mid-level managers have a unique leverage point: you are close enough to see how AI is already reshaping day-to-day work, yet positioned to influence upward strategy. Owning that space may be the difference between progress and paralysis.
4. Trust and Ambiguity Are the Real Battlegrounds
It’s surprising to see that in functions with the highest estimated AI potential (sales, marketing, software engineering, customer service) employee optimism is only moderate. Some may have tested AI tools and found the benefits overhyped. Others may fear job displacement.
Where stakes are highest, leaders need to manage not just implementation, but emotion—building trust while addressing realistic concerns. The way you frame uncertainty, the conversations you create, and the transparency you model will determine whether AI is seen as a threat or a partner.
5. The Skills Question Is Still Fuzzy
Employers don’t yet know:
How many AI-skilled workers they’ll need
What exact skills will be most critical
How quickly they can source them or retrain existing staff
What’s clear is that AI transformation requires a new level of leadership: vision-setting, cross-functional orchestration, and the ability to adapt strategy in real time.
5 Practical Moves for Mid-Level Leaders
Based on these observations, I believe the mid-level leaders have quite a unique position to drive the transformation, connect the dots and bridge the ambiguity.
Map the Value Chain
To mitigate the ROI challenge, the companies will need to know where to apply AI capabilities. While being close enough to the frontline, but detached enough from the weeds: Map clearly, where your team or function truly creates value. Start with processes and outcomes, don’t jump straight to “How can AI fit?” Instead, ask:
What’s driving most value and therefore what should we scale?
Where the costs occur now? And how can we frame the cost saving challenge?
Where are we most exposed to inefficiency?
Connect Across the Organization
Reach out to peers in other functions and tech teams. You’re uniquely positioned to translate business priorities into AI requirements—but only if you’re in the room early.
Open the Conversation with your Team
Make AI a standing agenda item within your teams. Share learning. Surface anxieties. Bring ethical and trust-related issues upward to the C-suite. Shape the narrative to be more nuanced and therefor real, so it’s not just “cost savings” or “shiny tech.”
Run Small, Value-Aligned Experiments
Build low-cost pilots around genuine business needs. Encourage your team to design solutions, not just react to them. Use hackathons or design sprints to get creative.
Define Metrics Now
Decide how you’ll measure AI’s impact in your area: efficiency, quality, risk reduction, customer experience. Defining success and KPIs will help you to prioritize the areas of AI inputs and see the impact early on.
Final Thoughts
AI is a disruption you can neither stop nor fully predict. But for mid-level leaders, it’s also an opening.
You don’t need to wait for the perfect roadmap. You can start small, act where you have influence, and shape the transformation from the ground up.
Check in with yourself: How do you personally respond to ambiguity? Can you step just far enough outside your comfort zone to take action, even if the destination isn’t fully clear?
Because the truth is, the comfort zone is shrinking—and those who lean in will help define what comes next.
Originally published on Substack.