How to redesign management for the agentic enterprise

AI has ignited a debate about how much management organizations need, and what managers will do as teams become smaller and more frontline work is automated. As intelligent systems and AI agents increasingly take on reporting, coordination, and other tasks that are repeatable and workflow-based, reducing management layers can appear to be a quick route to greater speed and efficiency. But the reality is, eliminating managerial work is not the same as eliminating the need for management.

While nearly three-quarters of companies plan to deploy agentic AI within two years, only 21% report having a mature governance model for autonomous agents.¹ At the same time, 42% of executives are concerned that employees could become overly dependent on AI for essential cognitive tasks.²

Some companies have already taken the leap to reduce their workforce and are beginning to backtrack. Of the one-third of surveyed leaders who let go of workers due to AI, more than half (57%) say they have come to question their decision.³

In a previous article, we explored how, as AI automates information gathering, status reporting, and other repeatable tasks, it raises the value of judgment, integration, and accountability. Management as we know it will also need to evolve: roles focused solely on supervision, approvals, and coordination are likely to diminish, while those that create context, connect people and intelligent agents, and guide decisions toward outcomes will become more consequential.

Strategy is set from the top down, but AI is changing how work gets done from the bottom up. Managers sit where those two forces meet. They are close enough to the work to see what’s changing, experienced enough to understand the broader consequences, and accountable for turning enterprise priorities into decisions and action. That makes the manager layer an essential link between strategy and execution, even as the nature of the role might change.

The value of managers shifts to orchestration

In organizations where AI is reshaping how work gets done, the team manager’s role increasingly becomes focused on orchestration. Managers have always connected strategy to execution: translating executive direction into action, communicating frontline realities back to leadership, handling escalations, coordinating across teams, and supporting employee performance. In an environment where work is distributed across people, AI agents, and automated workflows, these responsibilities become even more important.

Automating work does not automatically translate into greater efficiency or value. As AI speeds up analysis, reporting, and other parts of the work, operating cycles get faster. And that speed creates new coordination demands: more decisions to navigate, more trade-offs to resolve, and greater potential for activity in one part of the organization to affect another. Managers play a critical role in turning that increased speed into better outcomes.

AI also makes expertise more broadly available and allows more decisions to happen closer to the work. But access to more answers does not guarantee better outcomes. Someone still needs to interpret context, weigh competing priorities and risks, and understand implications beyond the immediate work. Managers provide that broader perspective.

This is why the design of human-AI interaction matters. Organizations that do this well (intentional human-AI interaction design) are nearly 2.5 times more likely to report better financial results and twice as likely to provide meaningful work.⁴ And business leaders see orchestration as a priority. Deloitte’s 2026 Human Capital Trends report found that dynamic work orchestration (the continuous, real-time coordination and reconfiguration of people, skills, data, and AI) is the most important trend for business leaders, with 88% rating it as extremely or very important.⁵

The takeaway is that value comes not from deploying AI alone, but from intentionally designing how people and technology work together. And managers are central to making that design work in practice.

Six shifts redefining the manager role

So what does that mean for how we think about the role of the manager layer in the organization? Some of what managers do today will take up less of their time. Other parts of their role grow. The six shifts below describe where managers’ value is moving as work becomes faster, more complex, and increasingly shared between people and AI.

1. Information becomes easier to aggregate but harder to interpret. Managers must create context and meaning for their teams and their leaders by interpreting patterns, explaining anomalies, and separating meaningful signals from expanding volumes of AI-generated analysis.  

2. Expertise becomes more accessible, making judgment the scarcer resource. Employees can now get expert-level answers from AI without going to their manager first. But they often lack the experience to know which answer fits and what to ask next. Managers help their teams build that skill: asking sharper questions, testing answers against what the organization needs, and knowing when to challenge the output.

3. Decisions increase in volume and move closer to the work. Faster work from AI means more decisions, but organizations can’t send them all up the chain anymore. More decisions now sit with managers, which makes their judgment more important. They decide which problems deserve effort, what gets escalated, and how local decisions fit enterprise priorities.

4. Learning must become more deliberate. When AI performs foundational tasks, employees may have fewer opportunities to learn through repetition and observation. Managers must deliberately create the apprenticeship experiences that the work once provided naturally, building judgment and the ability to assess AI-generated outputs critically.

5. Managers increasingly orchestrate systems rather than supervise static teams. Those systems may include employees, AI agents, platforms, workflows, external partners, and shared capabilities. Managers remain accountable for outcomes, but achieving outcomes requires determining the right blend of human and agentic effort, overseeing the development and performance of AI agents, and establishing clear oversight and fallback mechanisms where human judgment is critical.

6. Leading through change and uncertainty becomes a core managerial responsibility. As organizations redesign work around AI, managers become the critical link between the AI strategy and its execution: they help determine how work should change, guide their teams through the transition, and adjust ways of working as AI’s practical effects become clear. They are both leaders of the change and active participants in shaping it.

The risk of removing management layers

Managers have long filled an organizational gap: executives have an enterprise-wide perspective but are removed from many operational realities, while frontline employees understand the work but may lack the context or accountability to resolve complex organizational trade-offs.⁶ Managers connect these perspectives.

Organizations should still challenge legacy structures, and some managerial work, layers, and roles may no longer add value. However, organizations that focus exclusively on flattening structures risk solving yesterday’s efficiency problem while creating tomorrow’s coordination problem. If management layers are removed without considering the capability they provide, organizations risk creating a gap between strategy and execution. The unintended result is slower escalation, inconsistent decisions, unresolved tensions, and weakened accountability.

In organizations where managers today spend their time firefighting and compensating for structural weaknesses, AI may expose those weaknesses rather than eliminate them. AI will not automatically correct unclear governance, fragmented processes, or competing priorities.

AI changes the work, but it does not eliminate the need to connect strategy to execution, resolve trade-offs, build capability, and remain accountable for outcomes. Redesigning management therefore means redesigning around what organizations will continue to need (perhaps even more than they do today): judgment, accountability, learning, and orchestration.

Four moves for redesigning management layers  

Organizations can redesign management without having to choose between being stuck in the status quo and removing layers indiscriminately:

1. Assign clear decision accountability to each organizational layer relative to other management layers above or below. Specify which decisions managers own, which they influence, and which must be made elsewhere. Each layer should have a distinct decision-making purpose. Only humans can be accountable for decisions.  

2. Design explicitly for the interfaces between teams, workflows, and agents. Give managers the authority and incentives to coordinate resources beyond their formal teams, including shared agents, platforms, partners, and workflows. Assign who resolves competing requirements, and who is accountable for an end-to-end result.

3. Measure managerial load by outcomes instead of number of direct reports. Spans of control will become less useful as the primary measure of managerial burden. Begin with the business outcomes each layer must deliver.

4. Create governance mechanisms for translating bottom-up insight into enterprise decisions. Good strategy should work in both directions—bottom-up or top-down.

How Deloitte can help

Questions about management are becoming questions about organization design:

• What should managers be accountable for when AI can perform more of the work they once supervised?
• Where should decisions sit when expertise and intelligence are distributed across people and AI agents?
• How can organizations redesign management without creating a gap between strategy and execution?

Deloitte helps organizations answer these questions by redesigning management roles, decision structures, governance models, and operating models to improve performance in an AI-enabled environment.