Reskilling vs. replacement: Making the case for investing in people

It is nearly impossible to attend an industry conference, sit through a vendor pitch or read a white paper without hearing the same refrain: AI is going to reshape work, and employees will need to be reskilled or upskilled to stay relevant. The message is delivered with urgency, sometimes even with alarm. Consulting firms warn that those who do not act quickly will be left behind. Technology providers position their tools as essential to building the “AI-ready workforce.” Leaders declare that people are the most important asset in this new era.

On the face of it, the argument seems sensible. We are firmly in a period of rapid technological change, and employees will need different skills to succeed in roles that bear little resemblance to their job descriptions from even three years ago. Continuous learning and development (L&D) should be the logical answer. But the reality on the ground tells a different story.

Instead of building robust programmes to reskill their people, many organisations are opting for a shortcut. They are turning to the external market, choosing to hire employees who already have AI fluency, while simultaneously letting go of employees whose skills are tied to older models of work. It is faster, cleaner and easier to swap out talent than to take on the messy challenge of helping people evolve. Yet that decision comes at a cost — not just for the individuals who are displaced, but for the organisations that lose loyalty, trust, and institutional knowledge in the process.

This contradiction between the rhetoric of reskilling and the reality of replacement is becoming one of the defining challenges of the AI era.

The rise of AI layoffs

Recent months have made this painfully clear. At Accenture, 11,000 jobs were eliminated, with leadership warning that employees unable to move into AI-centric positions would be let go. Chegg laid off nearly a quarter of its workforce, citing the impact of generative AI (GenAI) on its core education business. Autodesk reduced 1,350 positions as it pivoted more aggressively toward cloud and AI. Even smaller organisations are following suit: Snorkel AI cut 13% of its team as it shifted to a new model. In May, Microsoft announced 6,000 layoffs, underscoring the widespread nature of the trend.

Perhaps the most striking case has been Salesforce. In early 2025, the organisation reduced more than 1,000 roles even as it was actively recruiting for AI-related positions. A few months later, CEO Marc Benioff confirmed that Salesforce had utilised its new Agentforce platform to automate vast portions of customer support, reducing the function from 9,000 employees to 5,000—a 45% reduction.

These examples are not anomalies. An analysis by Observer found that nearly 400 tech organisations announced layoffs in 2025, and many leaders explicitly pointed to AI or automation as the reason. The message to employees is unmistakable: If you do not already have the skills, your organisation may not wait for you to acquire them.

Rhetoric vs. resources

The paradox here is striking. Leaders stand on stages proclaiming that employees are their greatest asset and that reskilling will be the key to navigating the AI era. Yet, their actions often tell a different story. Reskilling is expensive, messy and time-consuming. It requires sustained funding, infrastructure and, most importantly, patience—patience for people to move through transitional roles where they are still learning and not yet operating at full speed. Those realities rarely align with quarterly earnings calls or cost-cutting imperatives.

It is far easier to hire someone who already has the right keywords on a resume. Post a job description, hire an “AI-ready” candidate and get back to business as usual. From a tactical perspective, the appeal is obvious. But it comes at a cost. Every external hire signals to current employees that their own growth is less valued. Over time, the result is not just a skills gap but a loyalty gap. Organisations lose not only institutional knowledge, but also the trust that underpins retention and engagement.

The not-so-quiet retreat of L&D

The contradiction becomes even clearer when looking at budgets. If reskilling were truly a strategic imperative, investment in L&D would be expanding. Instead, it is shrinking. The 2024 Training Industry Report found that the average budget for learning technologies dropped by nearly a quarter year over year. Projects once touted as transformative are now paused or cancelled. And in LinkedIn’s most recent Workplace Learning Report, 40% of L&D leaders cited budget constraints as their top challenge, even as executives continued to emphasise “future-proofing the workforce” as a strategic priority.

Employees are not blind to these contradictions. They hear the speeches about reskilling, then watch their training budgets disappear. What remains is scepticism and a growing sense that reskilling is more talking point than reality.

Jobs that don’t come back

It is also important to acknowledge that many of the jobs being cut will not be restored. Automation and AI do not simply replace tasks; they restructure work. When roles are redefined, they rarely return in their original form.

Evidence supports this. A Brookings analysis on labour displacement argues that retraining programmes have historically delivered mixed outcomes, with many employees unable to return to comparable roles after technology-driven disruptions. Research from Stanford’s Digital Economy Lab shows that younger employees in AI-exposed fields such as customer service, marketing and software support saw a 16% decline in employment between 2022 and 2025, even as overall employment grew. And MIT researchers estimate that AI could displace between 1.6-3.2 million jobs in the US over the next two decades, with many of those roles unlikely to return in their previous form.

The message is sobering. Even when new jobs emerge, they are not the same jobs. They often require different skills, structures, and expectations. That means displaced employees are rarely rehired one-to-one, unless organisations make a deliberate choice to build transition pathways. Without that, employees are left outside looking in, and the organisation loses valuable institutional knowledge that cannot easily be replaced.

Choosing to invest in people

So, what does real investment in people look like? It means weaving reskilling into transformation plans from the beginning, rather than tacking it on after roles have already been cut. Investing in people means treating learning as an ongoing process, supported through mentorship, rotations, and stretch assignments, rather than one-off workshops. It means holding leaders accountable for more than cost savings—tracking redeployment, internal mobility, and retention as real measures of success.

READ MORE: The augmented human: Why HR is the architect of our AI-powered future

It also means recognising the power of adjacent skills. Employees who already know your customers, processes and culture bring a foundation that cannot be bought on the open market. Pairing that institutional knowledge with targeted IA fluency can be far more effective than chasing understanding of the business. And, critically, it means protecting budgets for L&D. Cutting them may seem like an easy short-term fix, but it sends a clear signal that growth and transformation are not priorities.

The real differentiator

Investing in people is not an easy choice. It is slower, riskier and harder than hiring from outside. But avoiding it carries an even greater risk: a fractured workforce, a weaker culture and an organisation unprepared for the long-term demands of transformation.

Technology alone does not create sustainable advantage. People do.

AI should not be happening to us; it should happen with us. The organisations that thrive in the years ahead will not be the ones that adopt AI the fastest. They will be the ones who make their people central to the journey, building the future alongside their workforce rather than leaving them behind.

About the Author: Mary Faulkneris a principal with IA. This article was first published on HR Executive.