AI Transformation and the Illusion of Progress

The AI layoffs have begun. Across corporate America, workforces are getting smaller. Headcounts are getting capped. And entire functions are being reconsidered. Apparently, artificial intelligence is to blame.

The shift is particularly visible in Silicon Valley. Oracle laid off an astounding 30,000 people last week. At Block, the fintech company behind Square and Cash App, Jack Dorsey cut roughly 4,000 employees—about 40% of the company’s workforce. Dorsey argued that AI would allow a much smaller organization to operate effectively. At Shopify, teams asking for additional headcount are now expected to prove the task can’t be handled by AI. At Dropbox, roughly 500 positions were eliminated as the company began rebuilding itself around AI, with leadership acknowledging that some jobs would no longer exist in the future.

The trend isn’t limited to tech companies. At UPS, tens of thousands of jobs are being cut as the company automates sorting facilities and deploys AI-driven routing. At Nike, hundreds of corporate roles have been eliminated as the company leans further into AI-powered design, forecasting, and inventory systems. At Dow, thousands of jobs have been cut as automation reshapes its manufacturing operations.

All of this is being framed as artificial intelligence eliminating white-collar jobs. For now, many investors seem willing to accept that frame. AI is making work more efficient. AI is reducing the need for people. AI is allowing a smaller organization to do more. Taken together, it reads like the first large-scale labor-market consequence of the AI era. Yet that explanation doesn’t quite hold up.

AI as Helpful Cover

The scale of layoffs would suggest that companies have already captured significant productivity gains: AI has fundamentally changed the work, and fewer people are now required to do it. But the evidence indicates that this hasn’t happened yet. In one of the most rigorous studies to date, experienced software developers using AI tools actually took 19% longer to complete real-world tasks, largely because they had to spend time prompting, reviewing, and fixing AI-generated output.

Surveys of workers show a similar pattern: a substantial share of the time “saved” by AI is spent correcting mistakes or reworking low-quality output. This creates a false sense of productivity. Even inside tech companies, engineers report that AI increases workload intensity rather than reducing it, because outputs still require constant oversight. The evidence suggests the tools are improving, but haven’t reached the point where they explain the scale of workforce reductions we’re seeing.

It’s far more likely that we’re seeing a more immediate issue. Companies simply over-hired. During the pandemic and its immediate aftermath, hiring surged across industries. Between 2019 and 2022, Amazon doubled its workforce. Meta increased headcount by more than 60%. Shopify, Stripe, and others expanded aggressively in anticipation of sustained digital growth. That growth never fully materialized. When conditions normalized, those cost structures no longer made sense, and layoffs had to follow. The dynamic has little to do with AI. It’s a classic cycle of overexpansion and correction. This time, however, AI provides companies with a compelling narrative—one that shifts the story from “we hired too much” to “we’re building for what comes next.”

In many cases, there’s an even more uncomfortable explanation. The AI story is acting as cover for fundamental weaknesses in the business. At Nike, revenue has fallen 10% year-over-year, with declines across every geography. Net income dropped sharply as the company worked through excess inventory and slowing demand. Even now, the company is forecasting continued declines and describing its turnaround as “taking longer than expected.” UPS tells a similar story. Revenue has been under pressure for several years, with volumes declining as it unwinds major customer relationships and restructures its network. UPS has announced plans to cut tens of thousands of jobs as part of a broader effort to stabilize margins. In both cases, the layoffs coincide with weakening fundamentals: slowing growth, margin pressure, and the need to reset the business. AI may be part of the future plan, but it also provides a convenient way to frame decisions that are, at least in part, about fixing what isn’t working today.

One of the few places where we can clearly see AI’s impact on employment is in technology infrastructure. Companies like Amazon and Meta have found themselves in an infrastructure arms race. These companies have decided to lay people off and use the savings to buy more processors. And yet, even there, what they choose to cut may be an indication of where they’re correcting past mistakes. Facebook may have changed its name, but the metaverse didn’t show up on schedule. AI makes a great cover for that strategic misstep.

The Grand Illusion

What looks like a technology story is actually an important lesson about leadership in the modern age. It’s simply too easy to fool yourself into thinking you’re getting ready for the future when you’re really just dealing with the present or fixing the past. That may be fine for an external press release. But don’t confuse that with a future-focused strategy.

Spend time inside many organizations right now, and you can feel this tension playing out. Leaders are talking about transformation. They reference AI, market disruption, and what’s coming next. The language is forward-looking. But when decisions get made, something else takes over. There’s a quarter to close and targets to hit. The work that gets funded is the work that can be measured now. Often, that yields improvements in current efficiency, not preparation for what’s coming next.

That’s what makes this moment so tricky. Preparing for the future rarely shows up as efficiency in the present. It usually looks like the opposite. It requires investment before the payoff is clear. It pulls resources into areas that aren’t yet proven. And it creates friction inside an organization that’s built to optimize what already works.

When companies cut costs and call it transformation, they’re often doing something necessary. But they’re not necessarily doing something that expands their future. They signal change but then reinforce the existing model. Over time, that gap between current performance and future preparation begins to widen.

In reality, preparation for the future should make you less efficient in the short term, not more. If you buy flashlights and canned food to store up for an emergency, it’s going to add to your grocery bill this week, not reduce it. You should still do it so that your family can be ready for a big storm.

That’s the challenge with present-focused management. Most companies don’t get disrupted because they stand still. They get disrupted because they move quickly in a direction that no longer matters. Their leaders aren’t ignoring the future. They are simply talking about the future while reacting to the present.

Efficiency Isn’t Preparation

Future-focused leaders recognize that current efficiency isn’t preparation for the future. Those two things often get bundled together, especially in moments like this, when layoffs and AI spending can both be framed as “getting ready for the future.” But making the business more efficient today is not the same thing as making it more ready for tomorrow. One improves performance in the current model. The other expands the company’s ability to respond when that model starts to change. Leaders need to treat those as distinct agendas and resource them separately.

They also place small, deliberate bets. The future rarely arrives all at once. It begins as weak signals at the edge of the business—new technologies, new customer behaviors, new constraints that don’t yet look central. The goal isn’t to predict exactly which signal will matter most. It’s to make sure the organization is already learning, experimenting, and building familiarity before the shift becomes obvious. Small bets create options. They let a company move with more confidence when the world starts to tilt.

And they allocate resources against the future, even when it’s uncomfortable. This is the hardest part. The present will always make the stronger case. The numbers are clearer. The needs feel more urgent. The risks are easier to explain. The future asks for resources before the payoff is visible. But that is exactly what future-focused leadership requires: the willingness to put real time, money, and attention behind something that isn’t fully proven yet, because waiting for proof is often how companies end up arriving late.

When Larry Page returned as CEO of Google, he made a set of decisions that puzzled investors. He increased spending on administrative costs. The company invested more aggressively in long-term bets, many of which had uncertain payoffs. Entire parts of the business were reorganized around ideas that would take years to prove out.

The market reacted quickly. The stock dipped. Questions about Page’s discipline followed. From the outside, it looked like a company losing focus. From the inside, something else was happening. Page was positioning the company for what would come next. Things like a new mobile operating system, autonomous vehicles—and yes, AI. Google researchers published the transformer architecture that underpins modern large language models. Over time, those investments allowed Google to expand far beyond its original business and to participate in multiple waves of growth.

That kind of decision rarely looks good in the moment. The payoff is delayed. The logic is hard to explain. The results can’t be measured right away. Which is exactly why most organizations avoid it.

A Future-Focused Approach to AI

The AI moves that companies are making tell a clear story. AI is being deployed across functions. Organizations are getting leaner, and costs are coming down. From the outside, it looks like preparation for the future. What’s harder to see is whether those moves are actually expanding what the business can become.

A future-focused approach to AI starts somewhere different. It treats AI not just as a way to make current processes more efficient, but as a tool to explore the future of the business itself. Where could the model change? What work disappears, and what new work becomes possible? Which parts of the value chain should be rethought altogether, not just automated?

That means spending time in places where AI is still rough and incomplete. Prototyping new workflows, not just optimizing old ones. Reimagining how products are designed, how decisions get made, and how customers are served. The goal isn’t immediate efficiency. It’s learning—fast enough to see where the next version of the business might emerge.

You can see early versions of this in companies that are leaning into AI as a creative force, not just a cost lever. I recently wrote about how Eli Lilly has begun using AI to rethink drug discovery itself, compressing timelines and opening up entirely new approaches to finding treatments. AI isn’t just making existing processes faster. It’s changing what the company is capable of doing.

That’s the real choice in front of leaders. Use AI to do the same work with fewer people. Or use it to discover new work worth doing. Over time, those kinds of choices compound. One organization becomes more efficient at the business it already understands. Another begins to build the next version of its business while the first is still optimizing the last one.

The companies that matter in the next decade won’t be the ones that simply become more efficient versions of what they already are. They will be the ones who use that efficiency to move into something new. The question is whether the work being done today is creating new options or doubling down on existing choices.