Building the Agentic Enterprise, Part 9: The Human Side; Workforce, Roles, and Change

This is the ninth article in an 11-part series exploring what it takes to build an enterprise that runs on AI agents, not just AI tools. Each article examines a critical dimension of the journey and includes a "What It Takes" section with practical guidance for leaders navigating this transition.

---

The Dimension Organizations Underestimate Most

In Part 8, we covered governance, trust, and guardrails for agentic systems. But even the most robust governance framework will not deliver results if the people in your organization are not ready to work alongside agents. And for most organizations, they are not.

Talent readiness sits at just 20 percent across enterprises, the lowest score of any AI readiness dimension, well below technical infrastructure at 43 percent and data management at 40 percent. Despite 82 percent of enterprise leaders saying their organization provides some form of AI training, 59 percent still report an AI skills gap. IDC estimates that skills shortages will cost the global economy up to $5.5 trillion by 2026 in product delays, quality issues, missed revenue, and impaired competitiveness.

These numbers tell a clear story. Organizations are investing heavily in platforms, data infrastructure, and governance frameworks while underinvesting in the people who need to operate within them. People readiness is not a soft topic. It is the dimension that determines whether everything else in this series translates from strategy to practice.

Task Redistribution, Not Wholesale Replacement

The most persistent fear around AI agents is that they will replace workers at scale. The evidence so far points in a different direction. AI is reshaping jobs more than it is eliminating them. Over the next two to three years, 50 to 55 percent of jobs in the United States will be reshaped by AI, with most employees retaining the same or similar roles but facing substantially new expectations for how they work. The World Economic Forum projects 170 million new jobs will emerge by 2030 while 92 million will be displaced, a net gain of 78 million positions.

The more useful lens is task redistribution, not job replacement. Current data shows that 47 percent of work tasks across occupations are still performed solely by humans, 22 percent by technology, and 30 percent by a combination of both. By 2030, employers expect these shares to be nearly evenly split. Anthropic's January 2026 Economic Index found that 52 percent of AI usage was classified as augmentation, where AI helps humans work better, compared with 45 percent classified as automation, where AI handles the task independently. The balance tips toward augmentation, not replacement.

What this means in practice is that most roles will not disappear. They will decompose into component tasks, and those tasks will be redistributed. Some move to agents. Some remain with humans. Many involve both working together. The challenge for organizations is not whether to automate jobs but how to redesign work so that humans and agents each handle what they do best.

The Organizational Shape Is Changing

This task redistribution is changing how organizations are structured. Since AI agents can take on many entry-level tasks like data gathering, processing, and initial analysis, some organizations are finding that the traditional pyramid, with a broad base of junior workers, a middle management layer, and a senior leadership team, no longer matches how work gets done.

The emerging pattern is closer to a diamond shape: a smaller base of entry-level workers, a strengthened middle tier that trains, oversees, and manages agents alongside more complex work, and a leadership team focused on strategy and judgment. This does not mean eliminating entry-level positions. It means redefining what entry-level work looks like when agents handle the most routine components.

The implication for workforce planning is significant. If the entry point to your organization has historically been task-heavy, process-oriented work, and agents now handle much of that work, you need a new model for how people enter the organization, build skills, and advance. The apprenticeship path that many industries have relied on for decades needs to be redesigned for a world where the apprentice tasks are increasingly performed by agents.

Emerging Roles in the Agentic Enterprise

New categories of work are emerging as organizations deploy agents at scale. These are not hypothetical. They are appearing in job postings and organizational charts today.

Agent orchestration designers focus on the interface between humans and agents. They design the workflows, escalation paths, and interaction patterns that determine how agents coordinate with each other and with people. This role requires a blend of process design expertise, understanding of agent capabilities, and deep knowledge of the business domain where the agents operate.

Agent supervisors and operators monitor agent performance, handle escalations, and intervene when agents encounter situations outside their operating parameters. As we discussed in Part 5, the human-in-the-lead model depends on people who can set direction, adjust strategy, and exercise judgment that agents cannot provide. Agent supervisors are the operational expression of that model.

AI governance specialists establish and enforce the guardrails for agent behavior, auditing agent decisions and ensuring accountability. As Part 8 made clear, governance for autonomous systems requires ongoing attention, and someone needs to own that work day to day.

AI trainers and knowledge curators maintain the knowledge bases, context repositories, and feedback loops that agents depend on. As we covered in Part 7, the knowledge management dimension of data readiness requires people who understand both the business domain and how agents retrieve and use information.

By 2027, analysts project that half of all AI-enabled enterprise applications will require new oversight positions dedicated to governance, risk, and accountability. By 2026, 40 percent of G2000 job roles will involve direct interaction with AI systems. These are not distant forecasts. They describe the workforce redesign that is already underway.

Skills Evolution

The skills that matter are shifting. The AI talent gap has moved from "prompt engineering" to "agentic orchestration." Writing and maintaining code is becoming less of a differentiator as agents handle more of the implementation work. What is becoming more valuable are higher-order capabilities: systems thinking, judgment under ambiguity, cross-functional collaboration, and the ability to work effectively with AI tools as partners.

Human skills, including creative thinking, resilience, flexibility, and leadership, remain critical and are becoming more valuable precisely because they are the capabilities that agents cannot replicate. The 56 percent wage premium emerging for AI-fluent professionals reflects a market that is already pricing in this skills shift.

Yet most organizations are not preparing their people for this transition. Only 35 percent of leaders report having a mature, organization-wide AI upskilling program. Most training is fragmented, optional, and disconnected from how employees do their jobs. A 2026 Gallup survey of more than 22,000 employees found that only about 12 percent of workers report using AI daily, despite widespread enterprise deployment of AI tools. The gap between providing access and building capability is where most workforce readiness efforts stall.

The practical implication is that AI literacy cannot be treated as a one-time training event. It needs to be embedded in how people learn to do their jobs, integrated into onboarding, woven into performance development, and supported by hands-on practice in real work contexts.

Managing Hybrid Human-Agent Teams

Managing a team that includes both people and agents is a different discipline from managing either one alone. Leaders accustomed to directing people now need to direct work, deciding which tasks flow to humans, which to agents, and which involve both working together. The manager's role shifts from sole decision-maker to system architect: designing how work flows, monitoring outcomes, and adjusting the balance as conditions change.

This shift is revealing a leadership readiness gap. Only 22 percent of business leaders believe they can effectively manage teams that combine humans and AI agents. The organizational factors that determine AI's real impact, including culture, manager support, and talent practices, account for more than twice the influence of individual mindset and behavior. In other words, it is not enough to train individual employees on AI tools. The management layer needs to be rebuilt for a hybrid workforce.

Effective hybrid team management requires clarity about what agents can and cannot do, transparent communication about how work is being redistributed, and mechanisms for people to provide feedback on agent performance. It also requires leaders who can act as a stabilizing force during a transition that triggers real anxiety. Leaders who dismiss that fear rather than addressing it will find adoption stalling regardless of how good their technology is.

Change Management: The Make-or-Break Discipline

The organizations that succeed with agentic AI will not be the ones with the best technology. They will be the ones that manage the human transition most effectively. Change management for the agentic enterprise goes beyond traditional approaches because the change is continuous, not a one-time event. Agent capabilities evolve. Roles shift. The balance between human and agent work keeps adjusting.

Effective change management for this transition requires several elements. First, transparent communication about what is changing and why. People need to understand the business rationale for agent deployment, how their roles will evolve, and what support is available. Sugar-coating the implications or pretending nothing will change erodes trust faster than honest acknowledgment of uncertainty.

Second, practical training that connects to real work. Abstract AI literacy courses that teach concepts without application produce low engagement and lower retention. Training should be embedded in actual workflows, with people learning to work alongside agents in the context of tasks they perform every day.

Third, visible leadership commitment. When leaders use agents in their own work, talk openly about what they are learning, and demonstrate that they are navigating the same transition, it normalizes the change. When leaders delegate agent adoption to their teams while continuing to work the old way, it signals that the transformation is optional.

Fourth, mechanisms for feedback and course correction. People who feel they have no voice in how work is being redesigned will resist the change, and their resistance will be rational. Building channels for employees to surface problems, suggest improvements, and flag concerns converts potential resisters into participants.

Cultural Readiness: From Resistance to Adaptation

Culture is the invisible infrastructure that determines whether change management succeeds or fails. Research shows that 67 percent of organizations are culturally and operationally unprepared for AI transformation. The cultural barriers are often more stubborn than the technical ones.

Organizations with cultures that reward experimentation, tolerate productive failure, and empower individuals to try new approaches adapt to agentic AI faster. Organizations with cultures that punish mistakes, concentrate decision-making authority, and resist process changes struggle even when their technology investments are strong.

Building cultural readiness means creating psychological safety around AI adoption. People need to know that struggling with new tools is normal, that making mistakes while learning is acceptable, and that their value to the organization is not defined by tasks that agents can now handle. It means celebrating the people who find effective ways to work with agents and making early adopters into ambassadors rather than outliers.

The Opportunity Frame

Beneath the anxiety about displacement, there is a genuine opportunity that organizations should not understate. When agents handle the high-volume, repetitive, data-intensive work that consumes much of the average knowledge worker's day, people are freed to spend more time on the work that drew them to their careers in the first place: creative problem-solving, relationship building, strategic thinking, and the kind of judgment that comes from experience and empathy.

This is not aspirational rhetoric. Organizations that have deployed agents effectively report that employees spend less time on administrative tasks and more time on customer engagement, innovation, and cross-functional collaboration. The organizations that frame the transition as an opportunity to do more meaningful work, and follow through on that promise, see higher adoption and lower attrition than those that frame it purely as an efficiency play.

The promise must be genuine. If agents free people from routine work only to have that time consumed by more routine work or by layoffs, the trust deficit will undermine not just current deployments but future ones. The opportunity frame only works if the organization commits to reinvesting the freed capacity in ways that are visible and valuable to the people doing the work.

What It Takes: Workforce Readiness

This article maps to the workforce readiness dimension of the Agentic AI Readiness Assessment. Workforce readiness is the dimension organizations underestimate most, and it is the one that determines whether every other investment delivers its intended value.

Here is what readiness requires in practice:

Assess your AI literacy baseline honestly. Not whether you have training programs, but whether your people can work effectively with AI tools in their daily jobs. The gap between access and capability is where most organizations are stuck. If only 12 percent of your workforce uses AI daily despite having enterprise-wide access, you have a literacy problem, not a technology problem.

Plan for role evolution, not just role elimination. Map how each role in your organization will change as agents take on more tasks. Identify the new skills required, the tasks that will be redistributed, and the new roles that need to be created. Revisit this mapping regularly as agent capabilities and deployment scope change.

Invest in leadership readiness. Your managers will be managing hybrid human-agent teams. Most have no experience doing this. Leadership development programs need to include practical training on directing work across human and agent resources, managing the emotional dynamics of workforce transformation, and making decisions under ambiguity about how quickly to expand agent autonomy.

Build change management as a core capability, not a project. The agentic transition is not a change event with a start and end date. It is an ongoing evolution requiring continuous communication, iterative training, feedback mechanisms, and visible leadership engagement sustained over years, not months.

Design new career pathways. If agents are taking on entry-level tasks, you need new models for how people enter your organization, build skills, and advance. The apprenticeship model that works when juniors learn by doing routine tasks needs to be rethought when agents handle that routine work.

Measure adoption, not just deployment. The metric that matters is not how many agents you have deployed. It is how effectively your people work with them. Track daily active usage, time-to-competency for new agent tools, employee satisfaction with agent-assisted workflows, and the quality of outcomes produced by hybrid human-agent teams.

Up Next

In Part 10, we will turn to navigating the vendor landscape. With the readiness dimensions now mapped, from strategy and technology to data, governance, and people, the question becomes how to evaluate the vendors and platforms that will power your agentic enterprise. We will cover evaluation criteria, proof-of-concept design, reference architecture, and how to avoid the most common vendor traps.