If 2023 and 2024 were the years of AI hype cycles, 2025 is shaping up to be the year organizations have to actually deliver. But a new report from digital engineering consultancy Akkodis suggests enterprises are stuck on an awkward—but familiar—fault line: the people doing the work feel ready for AI, while the people responsible for scaling it aren’t so sure.
Titled The Capability Curve: Building the Next Generation Digital Enterprise, the report pulls data from 2,000+ business leaders (including 500 CTOs) and 37,500 employees across global industries. Together, they paint a surprisingly human story about AI transformation—a story less about flashy models or killer apps and more about skills, confidence, culture, and good old-fashioned organizational courage.
And if you’re wondering whether AI anxiety is still dominating water-cooler talk, it seems workers are actually feeling increasingly comfortable. Leadership? Not so much.
Workers Are Getting AI-Confident—Leadership Is Doing the Opposite
Akkodis highlights what it calls a “defining paradox”: 75% of workers now believe their leaders understand AI, up dramatically from 46% last year. Yet only 62% of leaders believe in their own AI implementation strategies, a significant 20-point drop.
That’s right—employees think their bosses “get” AI more than their bosses think they do.
If this sounds backwards, it’s because modern AI adoption has reversed old enterprise change-management logic. Tools like generative AI are introduced bottom-up and used in real workflows long before being made official. Workers experiment, iterate, and optimize. Meanwhile, leaders are left figuring out how to standardize, govern, and integrate those experiments across sprawling systems—without breaking things or creating compliance nightmares.
It’s not that leaders don’t believe AI works; they’re just not sure they can scale it responsibly. And that’s exactly where the confidence gap widens.
Skills: The Most Talked-About Problem Leaders Still Aren’t Solving
CTOs surveyed cite skills gaps as the number-one barrier to transformation. But here’s the twist: only 20% are using technology to actually track, measure, or support skill development.
If enterprises are serious about AI-driven transformation, tracking skills shouldn’t be optional—it should be foundational. Rivals in the consulting and talent ecosystem, such as Deloitte and PwC, have been loudly championing skills-based operating models for years. Akkodis’ findings suggest many organizations still haven’t operationalized that shift.
Put bluntly: companies say they don’t have the skills, but also aren’t measuring them.
Without structured skill visibility, leaders can’t match workforce strengths to emerging AI-powered workflows, nor can they design meaningful upskilling plans. Workers, meanwhile, continue experimenting on their own—feeding the paradox.
Productivity Gains Are Real—and Workers Are Spending the Time Wisely
One of the more optimistic insights in the report: employees say AI saves them around two hours per day, time they’re funneling into strategy, creativity, or more complex problem-solving. While “two hours per day” is the kind of round number that might trigger a raised eyebrow, it aligns with industry-wide productivity data from Microsoft, McKinsey, and Adobe.
The important signal isn’t the number—it’s the pattern. Workers aren’t just using AI to go faster; they’re using it to go bigger.
Leaders want to translate that pattern into repeatable, enterprise-wide productivity. But this requires data integration, process re-design, and system-level governance—not just licensing a chatbot for every employee and hoping for the best.
AI Might Reshape Teams, but Redeployment Beats Reductions
Another eyebrow-raising stat: 57% of CTOs expect AI to reduce workforce size within five years.
Cue the headlines—“robots are coming for your job.”
But that’s not the whole picture.
A slightly higher share—59%—say they plan to redeploy employees internally instead of eliminating roles. This aligns with a broader shift happening across HR tech and workforce strategy: replacing labor isn’t the goal; redeploying capability is.
Large enterprises are increasingly designing “skills-based mobility loops,” where workers shift laterally or diagonally toward higher-value tasks as automation absorbs repetitive work. AI isn’t replacing humans; it’s rewriting the value equation of human work.
The Playbook: Six Steps to Build an “AI-Confident Enterprise”
Akkodis distills the insights into six recommended actions. None are radical, but each reflects the evolving enterprise maturity curve:
-
Turn optimism into alignment – Bridge the employee–leadership confidence gap.
-
Redesign skills as a partnership – Make learning a two-way, tech-supported discipline.
-
Elevate AI as a leadership tool – Leaders should use AI, not just approve it.
-
Embed trust in hybrid workflows – Integrate AI with humans-in-the-loop at each step.
-
Scale systems with confidence – Move from experimentation to platform-level deployment.
-
Build a culture of shared accountability – Rethink how risk, data, and responsibility are shared.
This is effectively a blueprint for “responsible scaling”—the phase many enterprises find themselves stuck in today.
Case Studies: When AI Is More Than a Proof of Concept
The report includes several real-world case studies illustrating what “AI confident” looks like in practice:
-
Healthcare manufacturers using AI to optimize supply-demand decisions in seconds rather than hours.
-
Engineering teams connecting digital twins and model-based systems to bring product iterations and traceability into one connected loop.
These examples underscore a trend we’re seeing across the HR tech and enterprise tech ecosystem: real transformation happens when AI is paired with deep process engineering—not just added on top of existing workflows.
Digital twins, in particular, continue gaining traction in manufacturing, logistics, and automotive engineering. The shift toward model-based systems makes AI adoption more measurable, more governable, and more scalable—which is precisely where most leaders feel least confident today.
The Bigger Picture: AI Transformation Is Still More Human Than Technical
If there’s a single narrative thread in Akkodis’ report, it’s this: technology doesn’t transform enterprises—people do.
Workers are experimenting. Leaders are rethinking governance. CTOs are balancing ambition with operational risk. And everyone is searching for stable footing in an environment where AI evolves faster than enterprise structures can.
The companies that win the next decade won’t simply deploy more AI. They’ll build AI-confident cultures—where experimentation meets governance, where skills evolve continuously, and where humans stay firmly in the loop not because they have to, but because they make the system better.
In other words: AI’s real multiplier effect isn’t automation. It’s capability.
Join thousands of HR leaders who rely on HRTechEdge for the latest in workforce technology, AI-driven HR solutions, and strategic insights
