The Reskilling Trap

Part I: The Architecture of Failure

On a Tuesday morning in April, Jennifer Martinez sat in a small video call with her manager at a mid-sized financial services company in Toronto. She had been scrolling through an internal job listing for the past twenty minutes—a senior analyst role on a team she'd always wanted to join. The job posting read like it had been written for someone else: required skills included AI governance, advanced data analytics, and something called "agentic AI management."

Jennifer had been with the company for eight years. She'd learned every platform they owned, mastered their core data systems, and moved up from junior analyst to mid-level coordinator. On paper, she was exactly the kind of person the company should have groomed for the next level. But in the video call, her manager—who'd started in his role two years ago—told her something she wasn't expecting. "I think you'd need more upskilling first," he said. "We just don't have the tools right now to help you get there."

The call lasted four minutes.

Three weeks later, Jennifer accepted a position at a competitor. She took with her eight years of institutional knowledge, a network of internal relationships, and the kind of quiet expertise that usually gets underestimated on resumes. In the exit interview, she didn't mention the job listing or the conversation with her manager. She just said she was "looking for growth opportunities."

Her company is not alone. They're in the majority.

Across North America, organizations have spent an estimated $180 billion on workforce development in the past three years. They've licensed learning management systems, built career frameworks, sent out announcements about reskilling initiatives, and tied performance bonuses to skill development metrics. Human resources teams have never been busier. And yet, at the precise moment when the future of work demands more adaptability than ever before, employee engagement with that development infrastructure is collapsing.

Here's the paradox that defines 2026: 90 percent of organizations now use skills data to make hiring and promotion decisions. 93 percent track outcomes from their reskilling programs. And yet only 34 percent report that more than half their workforce is actually engaged in upskilling—while 31 percent see participation rates below 25 percent.

This isn't a problem with learning platforms. It's not a content issue. It's not even, fundamentally, about money.

It's a structural misalignment between what organizations are building and what people actually need to survive in a workplace that's changing faster than anyone expected.

The Participation Paradox: What the Infrastructure Hides

To understand what went wrong, you have to first understand what went right—or seemed to.

Starting around 2023, as companies began to absorb the shock of generative AI and the accelerating pace of technological change, a consensus began to form in executive suites and HR departments across the continent: the future workforce would be built on skills, not roles. Instead of hiring for jobs, companies would hire for capabilities. Instead of promoting based on tenure, they'd promote based on demonstrated competencies. It made sense. It was logical. It was, in theory, a way to democratize opportunity and make career progression less dependent on luck, connections, and the degree someone happened to earn twenty years ago.

So organizations invested. They built skills taxonomies—maps of every capability needed across the organization. They licensed platforms that could track skills development. They created internal mobility programs designed to help people move laterally into new roles. Some of the larger companies even built internal job marketplaces, designed to work like LinkedIn, except for positions within their own walls.

The infrastructure was sound. The intent was good. But something critical was missing: a reason for people to actually use it.

"We have this infrastructure in place. The learning management systems are there. The career pathways are mapped. The budgets are approved. But there's a fundamental disconnect between what organizations are measuring and what people are actually experiencing."

— Mohit Rajhans, AI Strategy Consultant, Think Start Inc.

That disconnect manifests in what researchers are now calling the "participation paradox." Organizations are tracking everything—logins, course completions, skill assessments—but the underlying behavior isn't changing. People are using the systems at the level of performance compliance, not genuine engagement. They take a required course because their manager mentioned it in a one-on-one. They log into the learning platform because they were included in a company-wide email. But they're not showing up because they see a clear path from their current life to a better one.

The data tells this story clearly. According to research from Fuel50, a talent mobility platform that works with several hundred organizations globally, only 34 percent of companies see more than half their workforce actively engaging with upskilling programs. The rest? They see minimal engagement. And here's the part that matters for the people involved: 94 percent of employees say they would stay longer at a company that invests in their learning. But the companies that invest the most aren't seeing higher retention. They're watching people leave anyway.

There's a reason for that mismatch. It's not that people don't want to learn. It's that learning disconnected from a visible career outcome reads as busywork—something the company is doing to itself, not something the company is doing for you.

The Psychology of Invisible Pathways

To understand this at a more granular level, you need to meet Marcus, though that's not his real name.

Marcus works in operations at a logistics company based in Atlanta. He's been there for five years, and he's good at what he does. He understands the workflow, he knows the systems, and he has relationships with people in other departments who trust him. When the company rolled out its new skills-based career framework last fall, Marcus was genuinely interested. He attended the launch webinar. He looked at the internal job marketplace and saw several roles that seemed like logical next steps: operations manager, operations coordinator in a different region, supply chain analyst.

The company had invested in a comprehensive reskilling program. There were online courses. There were certifications he could earn. There was a learning budget allocated to him each year. He decided to pursue the supply chain analyst role, which would represent both a lateral move and a slight bump up in pay—exactly the kind of internal mobility that the company's career framework was supposedly designed to enable.

So he signed up for online courses. He spent his evenings working through modules on data analytics and supply chain optimization. He downloaded the skills assessment tools and filled them out. And then—nothing happened. Six months later, he had completed the training. He had earned the certifications. But when he asked his manager about applying for an actual supply chain analyst position, the response was vague: "Those roles usually go to people with more background in that area," his manager said. Or: "We'd probably want to see you develop this skill a little more before you made that jump."

There were no internal analysts looking to transfer. There were no supply chain analyst roles actually open at his location. The job marketplace showed them as possible positions, but there were no actual openings, no actual hiring manager, no actual conversation.

Marcus stopped taking courses three months ago. He's now looking at external job postings. He figures if he's going to do the work to get the skills, he might as well get paid for them somewhere else.

This, too, is happening across the continent at scale. Organizations have built the infrastructure to track and validate skills, but they haven't fundamentally changed how they actually allocate work and opportunity. The career pathways exist on the learning platform. They don't exist in the actual company.

"The infrastructure exists. Almost nobody is using it. That's because most organizations designed learning programs when they needed to design career systems. A learning program serves the people who already know where they want to go. A career system reaches everyone, because it answers the question people are actually asking: What am I learning this for? And what changes for me if I do?"

— Mohit Rajhans, Think Start Inc.

The gap between having skills and having opportunities to use them is where the entire reskilling infrastructure breaks down. And in 2026, that gap is widening.

Part II: The Acceleration – AI as the Forcing Function

The reskilling problem existed before generative AI. But generative AI turned it into a crisis.

Last year, AI adoption across North American organizations sat at around 50 percent—a plateau it had held for several years. Then ChatGPT went mainstream, large language models proliferated, and organizations realized they either had to figure out how to integrate these tools or risk being left behind. By mid-2025, AI adoption had jumped to 65 percent. By now, it's approaching 80 percent in enterprises and 60 percent in mid-market companies.

That acceleration happened almost entirely ahead of any formal reskilling strategy.

Companies bought the tools first. They figured out what to do with them later. And the people who worked at those companies? They had to figure it out on the fly, usually without much support.

Keisha took the one-hour training session. She tried using the system. It was genuinely helpful about 60 percent of the time. The other 40 percent, the AI would generate responses that were technically accurate but missed crucial context about the customer's actual situation. So she'd edit them. And then she'd worry that her edits were being tracked, that somewhere in the company, someone was measuring how often she disagreed with the AI's suggestions, and that disagreement was being logged as "lack of trust in the system" or "resistance to automation."

She brought this up in a meeting with her manager. "I just want to understand," she said, "how this is going to change what I do here?" Her manager, who'd received the same one-hour training, didn't have a good answer. "I think," he said, "they're still figuring it out."

Three months later, Keisha started looking for a new job. She found one at a competitor—different company, similar role, but at a place that hadn't yet integrated AI tools. She's planning to leave next month.

What's happening to Keisha is one version of the reskilling failure. But there are others.

The Skills Expiration Crisis

Here's a number that should terrify HR departments and CFOs alike: professional skills now expire twice as fast as they did a decade ago.

Ten years ago, if you learned a skill in your mid-career, you could reasonably expect it to remain relevant for fifteen years. You might need a refresher. You might need to learn new applications of the skill. But the fundamental capability would hold value. Now? The effective shelf life of most technical skills is down to five years. And in fields like artificial intelligence, data analytics, and cybersecurity, it's down to three years or less.

This represents a fundamental change in the nature of work. It means that an employee who learned to do their job competently in 2023 might find that same job drastically changed by 2026. It means that the training you completed last year is increasingly likely to be partially obsolete by the time you finish it. And it means that the career trajectory people were promised—get good at something, get better at it, retire from it—is no longer available.

Fifty-three percent of organizations now report that critical skills in their industry become obsolete within three years or less. Fifteen percent say that window is under a year. This is not a technology-sector phenomenon anymore. It's happening in healthcare, manufacturing, financial services, and energy. It's happening across the economy.

The problem is that most learning infrastructure was built on the assumption that it wasn't.

"We still operate on an annual planning cycle. You design your training program in November or December. You deliver it from January through June. By then, the landscape has shifted. The skills you were training people on might have changed. New tools have emerged. The priorities have shifted. By the time people are actually competent at the new skill, it's halfway through its useful life."

— Emily Richi, Partner, Citrin Cooperman

Add to this the fact that most organizations still don't have robust skill assessments in place—only 53 percent even conduct formal skills assessments—and you end up with a situation where companies are essentially training people in the dark. They don't know what skills their people actually have. They don't know what's becoming obsolete. They don't know what the real gaps are. So they design training programs on assumptions. And by the time those programs are delivered, the assumptions have often changed.

The Manager Problem: Who's Supposed to Be in Charge Here?

If there's a single lever that determines whether reskilling actually works, it's the manager. Employees adopt skills systems 3 to 5 times faster when they see their direct manager actively using the system rather than just endorsing it. When a manager has regular career conversations with their direct reports—not once a year at review time, but monthly, or even bi-weekly—participation and actual skill development both increase.

The problem is that managers, in 2026, are overwhelmed.

They're dealing with uncertainty about AI's impact on their teams. They're managing burnout (52 percent of HR professionals are more concerned about burnout than IT leaders are, suggesting that traditional desk work is seeing higher stress levels than technical work). They're trying to fill roles that have become harder and harder to hire for. They're often navigating layoffs or restructurings. And they're being asked to be career architects on top of everything else.

This is the structural problem that most organizations haven't solved. They've built the infrastructure for managers to be career architects. They haven't actually given managers the time or support to do it. So the manager-as-accelerator model works great in theory, and it barely works at all in practice.

The result is that career conversations—the thing that actually drives people to engage with reskilling—remain rare. They happen sporadically. They happen when someone is already looking to leave. They happen when a manager is exceptional. But they don't happen systematically. They don't happen at scale. And because they don't happen, people make their own assumptions about what's possible, and those assumptions are usually pessimistic.

Part III: The Human Cost – What the Numbers Mean

So far, we've talked about this as an infrastructure problem. A systems problem. A problem with how companies have structured their reskilling initiatives. But the infrastructure problem has human consequences, and they're starting to be measured.

Seventy percent of organizations now struggle with both retention and skill obsolescence at the same time. These are two faces of the same failure: employees are worried about whether their skills are still relevant, they don't see a clear path to develop new ones, and so they leave. The replacement cost for that employee is between 10 and 20 percent of their annual salary. For a mid-market company losing 10 percent of its workforce to this dynamic, that's millions of dollars.

But the human cost is harder to quantify.

Seventy-two percent of organizations report that employees frequently express concern about their own skill relevance. In other words, more than two-thirds of the workforce is walking around with a low-level hum of anxiety about whether they're becoming obsolete. Meanwhile, 49 percent of employees say they're often exhausted and burned out. Four in ten report frequent feelings of anxiety and depression. Forty-two percent say they're often overwhelmed and unmotivated.

And here's the thing that most organizations still don't seem to understand: toxic workplace culture is over 10 times more predictive of attrition than compensation. People don't leave because the money is bad. They leave because they don't see growth, they don't see a future, and they feel like the organization is focused on everything except making them feel like they matter.

When you combine rapid skill obsolescence with limited reskilling opportunities and managers who don't have time to help people navigate their careers, what you create is a culture of stagnation—not because people aren't trying, but because the structure makes growth feel impossible.

Part IV: Where It's Actually Working – The Outliers

Not every organization is failing at this. A small number have figured out something different, and it's worth understanding what they're doing differently.

They also experienced something that most organizations don't measure: people starting to think of themselves as growing, even if they weren't formally taking courses. Because the conversation was happening, because managers were actively engaged in helping people think through their futures, the infrastructure felt alive instead of like busywork.

The other thing these companies did was make internal mobility visible and real. It's not enough to say that roles are available internally. You have to actually make those roles available. You have to back it up with hiring managers who are willing to take chances on people who've learned new skills but don't have formal experience yet. You have to create a culture where lateral moves are seen as growth, not as steps backward.

What all of these examples have in common is that they stopped treating reskilling as a learning initiative and started treating it as a career visibility initiative. The learning is important, but it's secondary. What's primary is making it clear that growth is possible, that it's intentional, and that the company is actually invested in helping people achieve it.

Part V: The Diversity Trap – Who Gets Left Behind

There's one more problem with the way most organizations have approached reskilling, and it's one that doesn't get talked about enough: the way it can systematically exclude people.

Early data suggests that younger workers—those aged 22 to 25 entering AI-exposed occupations—are showing a 14 percent drop in entry rates compared to prior years. That might sound like a small number, but the implication is significant: people are making the rational assessment that the skills they'll need to succeed are changing so rapidly that they might as well wait a few years before entering these fields. Or they're entering at companies that have clearer development paths. Or they're simply getting discouraged.

But the broader diversity issue is subtler. Reskilling programs, like all workplace programs, tend to be accessed by people who are already advantaged. People who have time to take evening courses. People whose managers are actively invested in their development. People who feel comfortable asking for help. People who aren't already stretched thin by caregiving, financial strain, or other life circumstances that make it hard to add another thing to your plate.

Last year, 43 percent of workers said they had trouble making ends meet. 26 percent had taken on an additional job or side hustle. 36 percent are caregivers. And yet reskilling programs are typically designed for people with spare time and headspace—designed, in other words, for people with privilege.

"We have these wonderful reskilling programs. But you can't really access them if you're working two jobs. You can't access them if you're providing childcare or eldercare. You can't access them if the cost of living is eating up your salary and you don't have the mental energy to take on anything else. And so the people who need the most support—the people most vulnerable to skill obsolescence—end up being the ones who can access it the least."

— Mohit Rajhans, Think Start Inc.

The manufacturing sector is grappling with this in real time. Manufacturing jobs are being automated, yes. But the jobs that remain are becoming more technical, more data-intensive, more digital. Which sounds like an opportunity for manufacturing to become more inclusive—if anyone can learn these skills, we're not gatekeeping on a four-year degree anymore. But in practice, it's often the opposite. The people who get access to upskilling and reskilling in manufacturing are already part of the existing structure. The people from outside—younger workers, workers from underrepresented groups, workers who've taken time out of the workforce—often find that the pathway is less clear.

"You need digital fluency. You need data literacy. You need adaptive leadership. These are all things that can be learned. But if you don't have access to quality training, if you don't have mentorship from people in the industry, if you don't have a manager who's invested in your development, you're going to struggle. And right now, those resources are unevenly distributed."

— Emily Richi, Citrin Cooperman

Part VI: The Path Forward – What Actually Needs to Change

If the problem is structural, then the solution has to be structural too. And it's not particularly complicated, at least in theory. The organizations that are actually making progress on reskilling are doing three things differently:

Beyond those three things, there are some specific levers that seem to matter:

The Human Piece: What People Actually Need

But here's the thing that no systems change can fully address: people need to feel like their employer cares about them as people, not just as repositories of skills to be managed.

"I hear this constantly: 70 percent of organizations say they're concerned about employee mental health, but only 27 percent offer an employee assistance program. Employers are embracing technologies at a breakneck pace, but they're not investing in the human infrastructure—the support systems, the mental health resources, the flexibility—that would actually help people adapt to that change."

— Mohit Rajhans, Think Start Inc.

When I asked Keisha Williams, the customer support agent we spoke with earlier, why she was leaving, her answer wasn't about the AI tools or the lack of training. "I just didn't feel like they cared," she said. "I felt like I was being asked to adapt to a machine, and I was being asked to do it with no support. I just wanted someone to say, 'Hey, this is hard, and we see that you're working to figure it out, and we're here to help.' Instead, I got a one-hour training session and an expectation to adapt immediately."

That's the human piece. That's what no infrastructure can replace.

Epilogue: Jennifer's Next Chapter

Three months after Jennifer Martinez left her job in Toronto, she's settled into her new role at a competitor. She's doing similar work—financial analysis, data management—but the company has a very different approach to career development. Her new manager sits down with her every month. They talk about what's exciting her, what skills she's developing, what she wants to learn next. When the company rolled out its skills-based career framework, Jennifer was asked what she thought of it. They asked her opinion. They adapted it based on feedback from people like her.

Is Jennifer's new company perfect? No. But they seem to understand something that her old company didn't: that investing in people's growth isn't a compliance activity or a strategic initiative on a PowerPoint deck. It's something that requires actual attention and care. It requires managers who show up. It requires organizations that mean what they say when they say they want their people to grow.

Jennifer is still taking courses. She's still developing skills. But now, instead of doing it in isolation, wondering if it will matter, she's doing it in conversation with people who seem to genuinely want to see her succeed. And that changes everything.

The Reckoning Ahead

We're at an inflection point. The World Economic Forum estimates that 59 percent of the global workforce will need reskilling by 2030. Korn Ferry projects 85 million unfilled jobs globally by 2030, representing $8.5 trillion in lost annual revenue. Board directors rank skills shortages as the top risk to organizational growth.

The infrastructure for addressing this exists. The money is being spent. The platforms are in place. But the gap between what we've built and what people are actually experiencing continues to widen.

Jennifer left. Keisha left. Marcus left. And they're not alone. The organizations that thought they were solving the reskilling crisis with better platforms and more training budgets are watching people walk out the door anyway.

The real reskilling crisis isn't about what people need to learn. It's about what organizations need to become—places where growth feels real, where managers are present, where learning is embedded into the work itself, and where someone at any stage of their career can imagine a future.

Until organizations solve that problem—not the systems problem, but the human problem—all the platforms in the world won't matter.