The New Management Model: Managing People, AI Agents and Organizational Capability
For more than a century, management has largely been organized around people and it is now changing
10/11/20265 min read
Managers were given teams. Organizational charts showed reporting relationships. Job descriptions divided responsibilities. Performance systems measured employees. Span of control was calculated by counting how many people reported to a manager.
Artificial intelligence is beginning to challenge that model.
The manager of the emerging organization may be responsible not only for employees, but also for AI agents, automated workflows, external specialists, organizational knowledge and increasingly complex decision systems.
This creates a fundamental management question:
What does it mean to manage when not everything being managed is human?
From AI Adoption to Organizational Redesign
The first phase of corporate AI adoption focused primarily on productivity.
Employees were encouraged to use AI to summarize documents, draft communications, analyze information, write code and perform routine administrative tasks.
That produced efficiency, but not necessarily organizational transformation.
Recent research increasingly points to the next stage: companies must redesign work itself.
McKinsey reported in September 2026 that 80% of surveyed organizations saw improvements in individual productivity from AI, while only 37% reported an EBIT impact. One explanation is what McKinsey describes as the "coordination tax": the enormous amount of organizational capacity consumed by handoffs, approvals, information gathering and coordination between functions.
The opportunity therefore extends beyond making individual employees faster.
Organizations can reconsider how work moves through the company.
Instead of asking:
How can AI help employees perform this process faster?
management can ask:
If we designed this process today, knowing that AI exists, would we design it the same way?
That is a much more consequential question.
AI Agents Are Entering the Organization
Generative AI primarily helped people create information.
Agentic AI can potentially perform sequences of activities.
An agent might gather information, analyze it, interact with software, prepare a recommendation, initiate a workflow or coordinate with other systems.
The organizational structure may therefore begin to look very different.
A manager might eventually oversee:
8 employees
15 AI agents
6 automated workflows
3 external specialists
Traditional span-of-control measures become much less informative.
The relevant management question is no longer simply:
How many people report to this manager?
It becomes:
What organizational capabilities, systems and decisions fall under this manager's authority?
Deloitte has recently argued that organizations entering the agentic era will need clearer ownership and decision rights around AI-enabled work.
That is important because an AI agent cannot be allowed to become an orphaned organizational capability.
Someone must own it.
From KYC to KYA: Know Your Agent
Financial institutions have spent decades developing Know Your Customer frameworks.
Organizations adopting autonomous AI may increasingly need another discipline:
Know Your Agent — KYA.
For every meaningful AI agent operating within an organization, management should be able to answer:
What is the agent?
Who owns it?
What is its objective?
What information can it access?
Which systems can it enter?
What actions can it perform?
What decisions can it make independently?
What financial or operational limits apply?
Can it communicate with other agents?
How is unusual behaviour detected?
Who reviews its performance?
Who can stop it?
Who is accountable when it makes a mistake?
The underlying principle is simple:
No authority without ownership.
An organization would rarely allow an employee to operate indefinitely without knowing who supervises that employee, what authority they possess or what systems they can access.
AI agents should not receive weaker governance merely because they are software.
Managers May Become More Important, Not Less
AI has produced widespread predictions about the disappearance of middle management.
Some management layers may certainly shrink.
But another possibility is emerging: the role of the manager becomes more important because the nature of management changes.
The future manager may need to be simultaneously:
people manager + change manager + AI orchestrator + capability developer + judgment layer.
AI can produce analysis.
It can generate alternatives.
It can identify patterns.
It can increasingly coordinate work.
But organizations still need someone to determine whether the analysis makes sense within the actual business context.
That requires judgment.
The New Management Skill: Managing Decisions
Traditional managers frequently spend considerable time gathering information, preparing reports and coordinating activities.
AI can reduce much of that work.
As a result, the manager's value may migrate toward something more difficult:
framing the problem → challenging assumptions → evaluating alternatives → understanding context → making trade-offs → accepting accountability.
The manager becomes less valuable because of access to information and more valuable because of the ability to determine what the information means and what should happen next.
This distinction becomes especially important as AI-generated analysis becomes abundant.
Organizations may soon face an unusual problem.
They will not suffer from too little analysis.
They may suffer from too much analysis and too little judgment.
The Risk of Losing Organizational Memory
Automation introduces another management risk that deserves more attention.
Experienced employees frequently know things that do not appear in procedures.
They know why an exception exists.
They remember what happened during the last crisis.
They understand which client requires additional attention.
They recognize when a technically correct answer does not make practical sense.
That is organizational memory.
If companies aggressively automate processes and remove experienced employees, they may inadvertently eliminate the contextual knowledge that allows organizations to handle unusual situations.
AI may know what normally happens.
An experienced employee may know why it happens that way.
Organizations therefore need to treat institutional knowledge as an asset that must be captured, transferred and maintained during AI transformation.
The Automation Trap
One of the easiest AI strategies to understand is:
automation → headcount reduction → cost savings.
It is also potentially one of the least imaginative.
Gartner recently predicted that some organizations replacing employees aggressively with AI will ultimately need to rehire portions of that workforce because institutional knowledge, talent pipelines and human capabilities were lost.
A more sophisticated model is:
AI → additional organizational capacity → redesigned work → employee development → innovation → growth.
The difference is important.
One model uses AI primarily to reduce the organization.
The other uses AI to increase what the organization is capable of doing.
Human Skills Become More Valuable
There is an interesting paradox in the AI transformation.
As machines become more capable, certain human capabilities may become more valuable rather than less.
These include:
judgment;
communication;
ethical reasoning;
influence;
negotiation;
conflict management;
contextual understanding;
creativity;
adaptability; and
accountability.
AI literacy will still matter.
But eventually AI literacy may become similar to digital literacy today: expected rather than exceptional.
The differentiating capability will be knowing when to trust AI, when to challenge it and when not to use it at all.
A New Definition of Management
Management has traditionally been described as achieving organizational objectives through people.
That definition may now be too narrow.
A more contemporary definition might be:
Management is the disciplined coordination of people, technology, knowledge and resources to achieve organizational objectives while maintaining clear accountability for decisions and outcomes.
That definition recognizes the organization that is beginning to emerge.
People remain central.
But they will increasingly work alongside AI systems capable of performing meaningful portions of organizational activity.
Five Questions Every Management Team Should Ask
Organizations do not need to wait for fully autonomous AI before preparing.
Management teams can begin with five questions:
1. What work should humans perform?
Identify activities where judgment, accountability, relationships or contextual understanding remain essential.
2. What work should AI perform?
Identify repetitive, analytical, information-intensive and scalable activities where automation creates genuine value.
3. Where should humans and AI work together?
Many of the strongest applications will combine machine capability with human judgment.
4. Who owns each AI capability?
Every significant agent, automated workflow and AI-enabled decision should have identifiable human ownership.
5. Where does final accountability remain?
Automation should never create ambiguity about who is responsible for an organizational outcome.
The Organization of the Future
The organizational chart of the future may not simply show people.
It may eventually show:
people → AI agents → automated processes → data → decision authority → accountability.
This will affect job descriptions, spans of control, operating models, performance management, governance structures and management education.
The organizations that succeed may not necessarily be those that automate the most.
They may be those that determine most intelligently what should be automated, what should remain human and how the two should work together.
And that brings management back to something remarkably traditional.
Technology can change how work is performed.
It can change organizational structures.
It can change how decisions are produced.
But someone must still decide what the organization is trying to accomplish, determine acceptable boundaries, allocate authority and accept responsibility for the outcome.
That remains management.
The tools are changing.
The responsibility is not. More about modern management can be found here.
© 2026 Applied Management Intelligence (AMI). All rights reserved.
Content, graphics, and materials are protected under international copyright law.
Reproduction, distribution, or use of this material without written permission from AMI is strictly prohibited.
