The uncomfortable truth about AI and work: PwC

Shebani Patel

Leader, Workforce Solutions, PwC US

Email

Dan Priest

Chief AI Officer, PwC US

Email

Key takeaways

AI transformation stalls when companies focus only on technology and ignore how work, workforce structures, and worker roles need to change together.

The biggest AI opportunity is not cost reduction. It’s redesigning roles, teams, and operating models so people and AI can create new forms of growth.

Companies pursuing advanced AI adoption need more than training programs. They need CHROs helping lead organizational redesign, incentives, and workforce strategy from the start.

There’s a question we’ve started asking CEOs that makes them pause.

You’ve told me what your AI can do. Now tell me specifically what your people do differently because of it. Not in theory—in practice. What does your underwriter’s Tuesday look like now? What decisions does your financial analyst make that they didn’t before? What did your marketing team stop doing, and what did they start doing instead?

Most can’t answer with any precision.

They can describe the technology. They may even cite the number of AI agents deployed, the processes automated, the pilots launched. But when asked to describe how the work has actually changed, the conversation gets vague fast.

"People are doing more strategic work." "They’re freed up to focus on higher-value activities." "We’re empowering them to work alongside AI."

These aren’t indications of a return on investment. They’re the same platitudes that we’ve been hearing for some time now. They ring hollow, because they don’t reflect how roles have changed, how you are structuring teams differently, or how you are measuring performance.

The problem is that most companies are stuck on technology and haven’t fully addressed the human side of the equation. As PwC US Senior Partner, Paul Griggs, recently shared, AI is creating a workforce dividend, a growing pool of human capacity, expertise, and potential. Deployed well, it becomes fuel for new markets, new products, and growth that wasn’t economically possible before. Captured narrowly, it becomes a line item in next quarter’s margin.

The deciding factor isn’t technology. It’s whether companies do the work to redesign how the workforce operates and how individual workers contribute. Until they do, the dividend will continue to elude them, and so will the growth it could fund.

Tackling only one-third of the problem

According to PwC’s 2026 AI Jobs Barometer, AI is both professionalizing and democratizing work. In some roles, it raises the premium on human expertise and judgment; in others, it makes specialized tasks more accessible to non-experts. The strongest growth is occurring in professionalized roles, where AI is making workers even more valuable.

While AI is changing people’s jobs, the results will vary significantly depending on how organizations respond. Yet the conversation is often dominated by two extremes: inflated claims about AI’s transformative impact and assumptions that AI’s primary effect will be workforce reduction. The reality is more nuanced, and far more dependent on the choices leaders make about work, workforce, and worker transformation.

To understand those choices more clearly, we use our 3W framework: Work, Workforce, and Worker.

Most AI transformations are designed around just one of these dimensions: the work. That covers what the technology can do, what processes it can automate, and what tasks it can accelerate. But it’s only one-third of the equation.

The workforce dimension looks at how the company organizes itself, from the level of small teams all the way up to divisions.

The worker dimension addresses what individual people do, what their days look like, and what they need to succeed.

Most companies are investing heavily in how AI can augment the first dimension—the work—while barely addressing the other two. But those two are where the AI dividend either comes to life or quietly dies.

The maturity mismatch

Many organizations are in the early stage of AI adoption, giving individuals generative AI tools to boost personal productivity. The people challenge at this stage is easy to manage: Train workers on the tools and let them experiment.

But the value most CEOs are chasing lives in the next two stages:

process and role automation, where AI agents transform entire workflows and the organization needs to fundamentally reinvent people’s roles; and

operating model reinvention, where the enterprise itself reorganizes itself around an AI-first architecture, with new service delivery models, operating in new markets, creating new products, new workforce configurations, and agentic systems working alongside humans.

Each stage demands exponentially more from the workforce and worker dimensions. Yet most companies’ people strategies haven’t evolved past stage one. They’re training individuals on tools and hoping the organizational transformation magically emerges from that.

One healthcare organization we work with is a good example of how AI maturity can evolve more effectively. In year one, the focus was appropriately at stage one: enabling AI use for various tasks by building awareness, developing comfort, and encouraging worker-led experimentation with AI tools.

As the organization is now moving into process and role automation through agentic AI, the people strategy has evolved to match. Leadership has identified the departments that they will be fundamentally reinventing, and for each, they’re addressing all three dimensions simultaneously: how the work changes, how the workforce is restructured, and how worker roles, skills, and opportunities expand. Most importantly, the organization is focused on leveraging newfound capacity to pursue new markets and perform work that wouldn't have been possible before.

The workforce dimension: more than headcount

Let’s address the elephant in the room. When leaders hear "workforce implications of AI," most think about headcount. That's the shallow version. If companies leverage AI efficiencies to fuel more layoffs, they may realize a short-term benefit, but they’ll squander a dividend that could be redirected toward customer growth, product innovation, market expansion, personalization, or new business models.

The real workforce question is a deeper structural one. People will not just disappear. But their daily lives will change significantly, and the value they contribute will change, too.

Reinvented roles. When AI takes over tasks within a role, you can’t just remove those tasks and hope for the best. If you want to realize long-term, strategic benefits, you need to deliberately recompose the role around new work, and you need to define those tasks as specifically as possible.

Here’s what this might look like in practice.

An underwriter whose week was dominated by data gathering and risk scoring now focuses on edge cases that the AI model flags, works on portfolio-level pattern recognition, and manages relationships with brokers.

A financial analyst who spent weeks building forecasts now interrogates AI-generated scenarios in real time.

A development team that spent 70% of its time on routine implementation now architects new features at a higher level of abstraction.

These roles aren’t smaller. They’re fundamentally different jobs that deliver more value. The reason most workers haven’t opted into AI transformation isn't fear of technology. It’s that nobody has shown them a credible, detailed picture of what their work looks like on the other side.

Team configuration. The traditional org chart has lots of junior people doing foundational work and fewer senior people reviewing it. When AI handles routine analytical work, that pyramid starts to collapse. What’s the right shape for the org chart when AI does much of what the base used to do? Most companies are still scratching the surface on this one.

Management layers. AI compresses information chains. Analysis that used to move through three levels of review can now surface directly to decision-makers. That doesn’t eliminate middle management, but it fundamentally changes its purpose, from information processing to judgment, coaching, and orchestration. Companies that don’t deliberately redefine these roles will find themselves carrying unneeded organizational weight that will slow them down when they most need speed.

Functional boundaries. When AI pulls insights across finance, operations, and customer data simultaneously, traditional silos look less like useful structures and more like artificial constraints. The strongest results are emerging from cross-functional teams organized around outcomes rather than disciplines.

Workforce architecture. The future isn’t just "employees plus AI." It’s a dynamic mix of permanent employees, contingent talent, and AI agents. The right configuration differs by function and process. Designing those processes deliberately rather than letting them emerge haphazardly is the difference between an operating model and a mess.

The worker dimension: ‘Be more strategic’ isn't an answer

If the workforce dimension is about organizational structure, the worker dimension is about what people need to thrive in that new reality.

As mentioned above, the strongest growth is occurring in professionalized roles, where AI increases the value of human expertise, judgment, and decision-making. Companies that want their workers to thrive should focus on designing roles that build on those strengths rather than simply using AI to automate existing tasks. Here’s how:

An honest deal. Workers are making a rational calculation about every new AI tool: Is this going to make my work more valuable, or will it hollow out my role until I’m redundant? The public conversation about AI has been dominated by fear-driven narratives. To counter that, the answer should be genuinely compelling: the boring, routine work will go to AI, and what will grow is the work requiring judgment, expertise, and creativity. This may be the honest answer, but people will only believe that when it shows up in how their company designs their role, how it measures their performance, and how it rewards them.

Skills that are defined precisely. "AI literacy" is too vague to act on. Recomposed roles demand specific capabilities, such as knowing when to trust an AI model’s output and when to override it. The new roles also require distinctly human skills that will become a bigger part of everyone’s job as AI absorbs the rote aspects: relationship building, interrogating model output rather than accepting it, creative problem-solving, and ethical reasoning.

Incentives that match the new reality. If you’ve redesigned the underwriter role around edge-case judgment but still measure how many policies they’re processing per day, you’ve designed the new work but incentivized the old work.

But measurement is only half of it. Compensation structures, promotion criteria, and recognition programs all send signals about what the organization actually values. Most organizations haven’t updated any of these to reflect AI-recomposed roles. When the underwriter who exercises exceptional judgment on complex edge cases gets the same bonus as one who processes volume, or when the developer who architects elegant systems sees no faster path to promotion than one who churns tickets, the message is clear regardless of what leadership says in town halls. People are rational. They follow incentives, not rhetoric.

Adoption that’s earned, not mandated. You won’t get people to adopt AI tools by developing better rollout plans or by mandating training hours. You’ll do it by creating clearly recomposed roles that people can see themselves in, providing specific skills development that builds their confidence, and giving them the agency to shape how AI shows up in their day-to-day work. When those elements are in place, adoption isn’t something you have to push. People will embrace it because the new way of working is visibly better than the old one.

Why all three dimensions must move together

We see many companies making the same series of mistakes. They invest in the work dimension, with new AI technologies, automation, and agents. Six months later, leadership realizes the organization isn’t structured to absorb the change. Six months after that, they realize people lack the skills or incentives for the new work. By then, twelve to eighteen months have passed.

Momentum has dissipated. The workforce has experienced the “transformation” as a series of disjointed, reactive interventions rather than a coherent vision, and they no longer trust the company or its AI plans.

Even when companies recognize the importance of workforce and worker reinvention, they may come late to the realization. One large technology company began its AI journey by experimenting with citizen-led AI tools and later investing in a suite of agents to augment core business processes. Only as they approached go-live did leadership confront the deeper question of how the work itself was changing and what that meant for the people doing it. What started as a technology initiative quickly expanded into a broader redesign of the operating model, aligning all three dimensions to support an agent-first future. They got there, but the sequential discovery cost them months they didn't need to lose.

Companies that get this right discover the transformation produces growth, not just efficiency. When you restructure the workforce and equip workers for AI-augmented roles, you unlock capabilities that weren't previously possible:

The underwriter evaluates risks in markets too complex to serve before.

The analyst surfaces opportunities quarterly forecasting would have missed.

The development team builds products in weeks that used to take months.

A focus on cutting costs with AI captures today’s margin but goes no further. Augmenting your people with AI uses today’s AI dividend to drive tomorrow’s revenue. But the most powerful growth play requires all three dimensions working in concert: work, workforce, and worker.

Who owns the missing two-thirds?

If the workforce and worker dimensions are where AI transformation succeeds or fails, there’s an obvious question most companies haven’t confronted: who in the C-suite actually owns them?

Today, AI strategy is typically led by some combination of the CEO, CTO, CIO, or chief digital officer. The CHRO is consulted to plan communications, design training programs, and manage workforce anxiety. That positioning reflects the old assumption that AI is fundamentally a technology initiative with people implications.

Let’s be direct about the consequence of that assumption: it’s a primary reason that transformations are stalling. Think about what the workforce and worker dimensions actually require: Organizational redesign. Recomposed roles. New team configurations and career architectures. Incentive realignment. Skills strategy tied to specific recomposed roles. A credible narrative that earns workforce opt-in.

All this is the CHRO’s (and his/her team) domain. And when the people who understand these disciplines aren’t at the table when foundational AI decisions are made, companies end up with a technology strategy that the workforce can’t absorb. Then they’re surprised when adoption stalls and call it a "change management problem." It’s not. It’s a governance problem. We’re missing the right people designing the transformation.

Our approach is to recognize that AI is fundamentally a people and organizational transformation, enabled by technology. Given that, the CHRO shouldn’t be a supporting player in the AI strategy; they should be co-authoring it.

For years, CHROs have been told they need "a seat at the table." The AI era demands more. It demands that the people strategy and the technology strategy are designed as one—and that the CHRO is one of the leaders holding the pen.

This is the CHRO's moment. The window to shape how AI transforms the workforce is open now. CHROs who are in as co-architects of the transformation will define the role for a generation. They will contribute to lasting advantages for their companies.

From Diagnostic to Action

What separates companies capturing value from companies that are stuck is the sophistication of their thinking about people and organizations.

So here’s the diagnostic every leadership team should sit with. If they can answer these questions honestly, and implement actions based on those answers, their organization will have a much more successful path to AI transformation.

Have you identified the growth plays this transformation should fund?

Can you describe—specifically, role by role—what your people's work looks like on the other side of this transformation?

Is your organizational structure designed for how AI-augmented work actually flows, or are you running new technology through old architecture?

Are your performance measures, incentives, and career pathways aligned to the recomposed roles, or are you still rewarding the work AI just took over?

Does your people strategy match the stage of AI maturity you’re pursuing, or are you applying stage-one thinking to stage-two and stage-three ambitions?