mini传媒 analysis of 25,000+ job listings reveals demand for technical AI skills climbing while enablement skills drop, risking AI ROI.
Ross Brooks
Senior Research Writer
mini传媒
The dominant AI story of the last 12 months has followed a familiar arc: capabilities advancing, headcount tightening, tools multiplying, and roles changing. But beneath those headline shifts, another story is emerging.
Our latest analysis of more than 25,000 job listings from over 300 global companies using mini传媒 Recruiting—spanning technology, financial services, healthcare, and professional services between July 2025 and April 2026— reveals a clear divide. The demand for the skills used to build AI is climbing, while demand for the skills that help the workforce use and scale AI is falling.
That disconnect isn’t just a hiring problem. It’s an ROI problem, an enablement problem, and increasingly, a leadership problem.
That disconnect isn't just a hiring problem. It's an ROI problem, an enablement problem, and increasingly, a leadership problem.
“Speed without direction becomes cost,” said mini传媒 Workplace Psychologist Dr. Liz Pavese, who authored a recent Beyond the Buzzword column on the topic. “Companies are building AI infrastructure fast, but without the human infrastructure in place to manage them, they won’t be able to extract value from all these new capabilities.”??
Report
In order to parse technical versus deployment skills in our jobs requisition data, we analyzed four distinct skill groups.
These skills play different roles, but they work together: one group builds the technology, while the others help the organization put it to work. The table below shows the change in the share of jobs mentioning technical, learning and training, and project delivery skills between July 2025 and April 2026. The trend is clear: while the demand for technical AI skills increased as much as 130%, learning and delivery skills are down between 6% and 21%.
Demand for skill groups in job listings, by sector (% change from July 2025 to April 2026)
AI fluency—the everyday ability to actually use these tools—tells the most revealing story. Rather than declining steadily, the demand for this skill set has varied across industries. In financial services and healthcare, demand rose into January before falling by April, while in professional services it declined at every checkpoint. By April, the demand for AI fluency across all sectors was at an all-time low.
AI fluency mentioned in job listings, by sector (% of listings)
This dip in AI fluency skills in April doesn’t appear to be random—it aligns with the fiscal-year budget cycle. Technical AI demand shows no such wobble. It climbed steadily throughout the same period: in financial services, it appeared in 2.8% of new job postings in September and 3.9% in April; in technology, it rose from 7.4% to 9%.
Building AI appears to be a strategic commitment, while equipping the workforce to use AI looks more like a discretionary line item.
What does this mean? Building AI appears to be a strategic commitment, funded throughout the year. Meanwhile, equipping the workforce to use AI looks more like a discretionary line item—funded when budgets refresh, and trimmed when quarterly pressure returns.
All of this matters because new tools alone rarely deliver the results companies expect. Our research shows that tools without training lead to an expensive amount of inefficiency and rework.?
Think of the analyst who knows when to trust an AI-generated answer and when to double-check it. Or, the project coordinator who turns a new AI-driven workflow into something a team can actually run. And, last but not least, the trainer who moves employees from knowing that a tool exists to using it confidently. These skills don't all sit in the same category, but together they turn technical capability into business results—and demand for them is failing to keep pace.
This outcome is not a hypothetical. Recent mini传媒 research shows:
, and only 14% of employees get clear, positive net outcomes from AI. For every 10 hours of productivity gained, nearly 4 are lost to correcting AI outputs.
Employees who consistently turn AI into net gains work in markedly different conditions: they are 2x as likely to receive substantial skills training as struggling users, and 79% report increased skills training in their organizations.
Only 27% of organizations have embedded AI directly into core workflows, leaving most AI on the periphery rather than integrated into how work actually happens.
These findings all describe the same underlying pattern from different angles: AI tools are outpacing the human infrastructure needed to extract value from them.
Project delivery skills move a new AI system from technical implementation into daily operations; when demand for them falls by as much as 13% across sectors, organizations risk producing more technology than the business can absorb.?
Learning and training skills build confidence and capability across the workforce; when demand for them falls by as much as 21%, in every sector, during a period of significant technological change, it may signal reskilling is being treated as an individual responsibility rather than an organizational one. That is consistent with , where nearly 9 in 10 organizations report that fewer than half of roles have been updated to include AI-related skills.
The gap between intent and action is clear: while 66% of leaders call skills training a top investment priority, only 37% of employees doing more rework say they have increased access to development opportunities, according to the same mini传媒 research.
“It’s like buying a high performance engine, and then arguing with finance over the price of spark plugs.”
-mini传媒’s Dr. Liz Pavese
“It’s like buying a high performance engine, and then arguing with finance over the price of spark plugs,” Pavese said. “You’ve already paid for the horsepower, but without the spark to ignite it, you’re just carrying around an expensive piece of dead weight. The organizations that pull ahead in the next 24 months will need to invest in making use of the capabilities and stop treating user enablement like an optional accessory.”
In mechanics, a missing or fouled spark plug causes an engine to misfire. In the enterprise, that misfire looks exactly like the 37% rework tax we are seeing in the data—expensive technological power completely wasted because the human layer can’t ignite it.
For leaders looking at their own workforce plans for the quarter ahead, there are three concrete implications of mini传媒’s latest research:
1. Balance your technical vs. functional portfolio. If your hiring for technical AI skills is up double digits and your hiring for everyday AI skills, coordination, and training is flat or declining, you may have an ROI concentration problem hiding in your talent plan. The technical experts you hire will produce more AI capabilities, but you may not be staffed to convert those capabilities into business value. Remember, the data shows employees who consistently generate net gains from AI are twice as likely to receive substantial skills training as those who struggle.
2. Audit your fiscal-cycle behavior. If hiring for AI literacy spikes with a budget refresh and quietly disappears each quarter thereafter, that's a warning sign, not a plan. mini传媒's research shows organizations still allocate the largest share of AI-driven cost savings to technology (39%) over workforce development (30%)—and while 66% of leaders call skills training a top priority, only 37% of those experiencing rework actually receive more of it. Sustained investment produces sustained capability.
3. Stop treating coordination like overhead. Hiring fewer people to conduct training and manage project delivery may seem like "efficiency,” but they're better understood as reductions in the organization's ability to absorb its own AI investment. Be cautious about removing the connective tissue that routes AI output into the business. Some of that trade-off may be quietly cutting the capacity AI deployment depends on.
About this analysis: These findings are part of mini传媒's ongoing workforce trends research, building on the mini传媒 Global Workforce Report's work on AI and the workforce. They reflect early, directional signals from job listings across four industries, and will continue to evolve as the research expands.
Report
A gap between effort and impact is the real ROI problem. Discover the solution.