Rethinking your workforce for the agentic AI era

If your organisation is like many, it’s become increasingly specialised: deeper functions, taller org charts, and narrower roles. Now, that model is shifting. In our 2026 AI Business Predictions, we call this shift the 'rise of the generalist:' a move toward broader, outcome-focused roles, which is already underway—we’re seeing it across industries, in back offices, and frontline teams. It’s one of the reasons why AI can make people—if they have the right skills and are in the right roles—more valuable than ever. 

In software development, for example, you may no longer need teams of people at work in each stage: generating solution architecture and user stories, creating test cases and test scripts, or troubleshooting, reviewing, and documenting. One experienced software engineer can orchestrate teams of AI agents at work in many or even all stages. 

As they shift from narrow execution to broader responsibility, these engineers—and other specialised, high-value employees like them—can become generalists: working across more processes, making faster decisions, and focusing on bottom-line impact. 

The new generalists don’t build themselves. They’re the result of a pipeline that takes AI-literate, early-career workers and gives them—often through an apprenticeship model—specialised, real-world experience. That software engineer can work over the full development cycle because they’ve already built software without AI agents’ help. And it’s not just software. If you don’t build your pipeline now, you may soon find yourself without the expertise and leadership needed to correct AI-generated errors, spot systemic risk, or seize new market opportunities. 

There’s another likely benefit to early-career workers: When they’re equipped with AI agents, they can meaningfully contribute faster than before. At PwC, for example, our audit teams use AI to help execute many specialised audit tasks and deliver a more seamless client experience. So, instead of focusing our entry-level auditors on narrow technical tasks, we also teach them to use AI agents—and to think critically, support our longstanding focus on independence, audit quality, and data security, and focus on clients’ big-picture needs. 

If your organisation is like most, it’s structured like a pyramid. A small leadership team sits at the top, supported by a larger middle tier of managers and highly trained specialists. Below them is an even larger group of employees, many of whom are recent hires, who focus on routine manual or process work. But since AI agents can take on many entry-level tasks—such as data gathering, processing, and reporting—some companies are reducing the pyramid’s 'base.' They’re also repurposing the middle tier to train, oversee, and manage agents. The result is a diamond-shaped structure: a small leadership team, a strong middle layer, and a narrow base of new talent. 

For many companies, this diamond structure can be a good choice. It can help you manage AI agents at scale, make operations nimbler, and focus people on higher-value work. But it should be a choice, not a default, because for many companies, this diamond structure can carry medium- and long-term risks. 

If your company is based on knowledge work, reducing early-career workers could limit apprenticeship and starve your organisation of its future: You won't be developing the high-performing generalists and business leaders (with experience in your unique processes) who could fill your upper ranks and be your differential. For these knowledge-intensive firms, an hourglass structure—which we’re building in-house at PwC—can be a better choice. Here, entry-level roles can expand as AI-literate, business-savvy, and change-ready employees ramp up quickly, so they can contribute at a high level across workflows. Experienced specialists can expand their reach, combining their creativity, experience, and strategic insight with the scale and speed of AI agents. 

With fewer routine tasks performed by people who need routine management, there’s likely less need for expensive and often change-resistant middle managers. Instead, these experienced workers can take on higher-skill roles in exception handling, coaching, and high-value decision-making. Leaders can gain broader visibility and greater influence, as AI-enabled workflows accelerate the need for new operating and business models. With a strong, capable base and a forward-looking leadership tier, connected by a lean, high-performing middle, your org chart looks like an hourglass. 

The rise of AI-enabled generalists is already changing how core business functions operate. In finance, IT, marketing, and HR, we’re seeing real examples of how agentic workflows are expanding roles, elevating responsibilities, and helping teams focus on outcomes instead of tasks. These changes aren’t uniform. Each function is evolving in its own way, and leading organisations are designing roles and structures to keep pace. 

The organisations that are successful with AI typically aren’t starting at the edges. They’re starting at the top. Leadership identifies a few high-impact areas for investment, sets a clear strategy, and guides a focused path to value. That can include building a user-friendly orchestration layer to help create and manage workflows and focusing on Responsible AI to enhance performance and grow stakeholder trust. 

But strategy alone likely won't deliver outcomes. Here’s how you can get started. 

• Redesign workflows—and think big. Instead of asking how AI can improve a process, ask how you can reinvent it with AI at the core. People should be at the helm: always in charge, always making the big decisions, and always accountable for outcomes. But AI agents may be able to shave off as much as 50% of human effort. 
• Map out the impact. As you construct or assess an AI business case, make sure that it covers which roles are likely to be automated, augmented, or created—and what these changes mean for headcount, career paths, and reskilling. 
• Clarify who does what. For each role in an AI-enabled workflow, classify the work into AI-only / human + AI / human-only, then rewrite role purpose and skills accordingly. If you don’t have active AI roles yet, take this decoupling and tagging step before proceeding. 
• Transform key positions. For critical roles, create an explicit 'AI collaboration profile' that defines how AI should be used, what should remain human, and which skills matter most. Consider this data for talent reviews and workforce investment decisions. 
• Drive strategic alignment. Work with fellow leaders to modernise the job catalog, define AI-era roles, and standardise AI responsibilities and skills. As AI-enabled generalists spread, cross-functional standards and collaboration can become a strategic imperative. 
• Double-down on human intelligence. In the age of AI, people are still your differential. Hire and develop AI-savvy, change-ready, early-career workers; show current employees what their future roles might look like and how they can prepare; reward those people who strive to create advantage with AI; and cut through any organisational 'permafrost' that might get in their way.