Your Org Chart Was Designed for Human Information Flow

AI Strategy Series | Article 13 of 20

Building the Intelligent Enterprise

For more than a century, organizations have designed their management structures around a fundamental constraint: information must move between people before coordinated decisions can be made.

Employees collect information. Supervisors consolidate it. Department managers interpret it. Executives evaluate competing priorities. Decisions move back through the organization, where employees translate direction into action.

The familiar organizational chart reflects this arrangement.

It establishes reporting relationships, management responsibilities, functional boundaries, and decision authority.

But it also describes something less obvious.

The organizational chart is an architecture for moving information through the enterprise.

That architecture emerged when collecting, interpreting, distributing, and coordinating information required substantial human effort.

Artificial intelligence is changing those economics.

AI systems can retrieve information from multiple enterprise applications, reconcile records, summarize complex situations, identify patterns, and prepare recommendations without requiring every intermediate human handoff.

AI agents can extend those capabilities by coordinating activities and executing authorized actions across business functions.

This raises a strategic question that extends well beyond automation.

If AI fundamentally changes how information moves through an organization, why should we assume that the organizational structures built around traditional information flows will remain optimal?

The answer is not that organizational hierarchies will disappear.

They will not.

Enterprises still require leadership, accountability, professional expertise, managerial judgment, and clearly defined authority.

But some organizational structures exist because information was historically expensive to collect and coordinate.

As AI reduces portions of those costs, the economic justification for certain management layers, departmental handoffs, reporting relationships, and coordination mechanisms may change.

The enterprise must distinguish between organizational structures that create genuine management value and those that primarily compensate for historical limitations in information processing.

The AI-native enterprise will not necessarily eliminate hierarchy. It will redesign hierarchy around decisions, accountability, and outcomes rather than the manual movement of information.

The Organizational Chart Is More Than a Reporting Structure

Most executives think of an organizational chart as a representation of authority.

It identifies who reports to whom.

It defines departments.

It establishes managerial spans of control.

It provides a framework for assigning responsibility.

These functions remain important.

But organizational structures also solve an information-processing problem.

Consider a conventional enterprise with several hundred employees.

Frontline workers observe operational conditions.

Supervisors collect and interpret their observations.

Department managers consolidate information into functional reports.

Senior executives compare those reports against enterprise objectives.

Decisions are communicated through established reporting relationships.

Each management layer performs some combination of information aggregation, interpretation, coordination, prioritization, and authorization.

This arrangement allows a large organization to operate without requiring every employee to communicate directly with every other employee.

It reduces complexity by establishing defined channels.

But it introduces costs.

Information may be delayed.

Important context may disappear during summarization.

Departments may interpret the same facts differently.

Decisions may wait for scheduled meetings.

Employees may receive direction after the conditions that prompted it have changed.

The organization accepts these limitations because the hierarchy provides coordination, specialization, and accountability.

Historically, there have been few practical alternatives.

AI changes that calculation.

It can reduce the effort required to collect and interpret information across organizational boundaries.

That does not eliminate the need for management.

It changes the information-processing work management must perform.

Why Hierarchies Emerged

Organizational hierarchies did not develop arbitrarily.

They reflect economic and operational realities.

In The Nature of the Firm, economist Ronald Coase explained that firms exist partly because coordinating activities through internal management can be less expensive than coordinating every transaction through external markets.

Oliver Williamson later expanded the analysis of organizational governance and transaction costs.

Within organizations, hierarchy provides a mechanism for coordinating activities, resolving conflicts, assigning resources, and exercising authority.

Management structures also address limits on human attention.

A supervisor cannot effectively evaluate an unlimited number of employees, activities, and decisions.

Organizations therefore establish manageable spans of control.

As enterprises grow, additional management layers may become necessary to coordinate increasing complexity.

Functional departments create specialization.

Reporting relationships establish accountability.

Policies standardize decisions.

Management processes reconcile competing priorities.

These structures are not inherently inefficient.

They are responses to the cost of coordinating work and information.

The important point is that organizational design depends partly on the capabilities and limitations of the technologies available.

When information is difficult to collect, organizations create reporting mechanisms.

When coordination requires extensive human communication, organizations establish intermediary roles.

When decisions depend on information distributed across functions, organizations create committees, management layers, and escalation processes.

As the cost of information processing changes, the economics supporting some organizational structures change with it.

That does not mean technology alone determines organizational design.

Human judgment, organizational culture, regulatory requirements, and accountability remain essential.

But ignoring the effect of AI on coordination costs would be a strategic mistake.

The Five Information Functions Embedded in Management

To understand how AI may affect organizational design, we should examine the information-processing activities performed by management.

I would distinguish five functions.

1. Information Collection

Managers gather information about operational performance, customer conditions, employee activities, financial results, and emerging problems.

Historically, this requires reports, meetings, interviews, spreadsheets, and manual consolidation.

AI can automate portions of this collection by retrieving information from authoritative enterprise systems.

2. Information Interpretation

Managers analyze information to identify patterns, anomalies, risks, and opportunities.

AI can assist with interpretation by detecting relationships, preparing explanations, comparing alternatives, and identifying conditions requiring attention.

3. Information Distribution

Managers communicate information to employees, other departments, and executive leadership.

AI can distribute relevant information directly to authorized recipients and generate context-specific summaries.

4. Decision Coordination

Managers reconcile competing requirements, evaluate tradeoffs, and coordinate decisions across teams.

AI can prepare alternatives, identify dependencies, and support coordination, although consequential decisions may continue to require human judgment and authorization.

5. Accountability and Leadership

Managers establish direction, exercise legitimate authority, develop employees, resolve conflicts, and accept responsibility for outcomes.

AI may support these responsibilities.

It does not assume the organizational accountability associated with them.

This distinction is critical.

The first four functions contain activities that AI may significantly accelerate or partially automate.

The fifth represents responsibilities that cannot simply be transferred to software.

The organizational implications become clearer when these functions are separated.

If a management position exists primarily to collect, consolidate, and distribute information, AI may substantially change the work associated with that position.

If the position exists to exercise judgment, develop people, resolve complex conflicts, and remain accountable for outcomes, its value may become more important.

The question is not whether managers are necessary.

It is which managerial activities will continue to justify the organization’s investment in management capacity.

The Information Bottleneck Is Moving

Traditional enterprises frequently experience information bottlenecks.

An operational problem occurs.

An employee identifies it.

A supervisor receives the information.

A manager evaluates its significance.

Another department is consulted.

A decision is escalated.

Approval is granted.

The action returns through the organizational structure.

The elapsed time may be substantially greater than the time required to execute the actual work.

Consider a manufacturing organization experiencing repeated equipment failures.

Maintenance technicians document incidents.

Supervisors consolidate maintenance records.

Reliability engineers analyze failure patterns.

Operations managers evaluate production consequences.

Procurement reviews replacement-part availability.

Finance evaluates expenditure requirements.

Senior management authorizes a corrective investment.

Each function contributes legitimate expertise.

But the process may contain considerable information-transfer overhead.

An AI-enabled architecture could retrieve maintenance histories, identify recurring failure patterns, correlate operational conditions, evaluate inventory availability, and prepare a consolidated assessment.

The information could be presented simultaneously to the relevant decision-makers.

The organization would still require qualified engineering judgment, financial authorization, and operational accountability.

But several intermediate reporting activities could become unnecessary.

The bottleneck moves.

Instead of waiting for information to be assembled, the enterprise may be waiting for a consequential decision.

That is an important shift.

When information becomes available faster, decision authority becomes a more visible constraint on enterprise performance.

Organizations that accelerate information processing without redesigning decision rights may discover that their management structures become the primary obstacle to realizing AI’s benefits.

Faster Information Does Not Automatically Mean Better Decisions

The ability to produce information quickly is not equivalent to the ability to make sound decisions.

AI can generate recommendations at a pace that exceeds human capacity to evaluate them.

An executive who previously received a weekly operational report may now receive continuous alerts, forecasts, and proposed interventions.

A procurement manager may receive hundreds of AI-generated recommendations.

A security leader may receive an expanding stream of prioritized findings.

A finance executive may receive continuously updated scenario analyses.

The information bottleneck has not necessarily disappeared.

It may have shifted from information production to human attention.

This creates a new organizational challenge.

The enterprise must determine which information requires human attention, which decisions can be delegated, and which conditions warrant escalation.

Without that discipline, AI may increase the volume of management information without improving management effectiveness.

The result could be more dashboards, more notifications, more recommendations, and more meetings.

That is not transformation.

It is information overload with better technology.

The solution requires deliberate decision architecture.

For each significant decision category, the enterprise should establish the relevant authority, information requirements, escalation conditions, and expected outcome.

AI should then support that structure.

The objective is not to deliver more information to more managers. It is to deliver the right information to the appropriate decision authority when intervention creates value.

Why Functional Silos May Become More Expensive to Preserve

Functional departments provide important benefits.

They concentrate expertise.

They establish professional standards.

They support workforce development.

They create clear responsibility for specialized activities.

Finance, legal, procurement, human resources, operations, and IT perform distinct functions for legitimate reasons.

AI does not eliminate those distinctions.

But functional boundaries often introduce coordination costs that extend beyond the expertise they preserve.

Consider a customer onboarding process.

Sales collects customer information.

Legal reviews contractual terms.

Finance establishes billing arrangements.

Operations prepares service delivery.

IT configures systems.

Customer success assumes responsibility for the ongoing relationship.

The customer experiences one business process.

The enterprise manages it through several departments.

Each handoff may require additional documentation, communication, verification, and approval.

Some of these activities are necessary.

Others exist because systems and departments historically lacked a shared, reliable view of the process.

AI can potentially reduce this fragmentation by assembling information across authorized systems and coordinating activities through common workflows.

But that capability creates a strategic choice.

The organization can retain the original departmental process and use AI to accelerate individual handoffs.

Or it can redesign the workflow around the end-to-end customer outcome.

The second approach may produce substantially greater value.

It may also require changes in process ownership, performance measures, management responsibilities, and decision authority.

This is where organizational redesign becomes more consequential than technology deployment.

AI may make departmental coordination cheaper, but enterprises must decide whether their existing departmental boundaries still represent the most effective way to organize work.

The Difference Between Functional Expertise and Functional Fragmentation

Organizations should avoid treating every departmental boundary as an inefficiency.

Specialization remains valuable.

A legal professional’s judgment is not interchangeable with a procurement analyst’s expertise.

A clinical supervisor’s responsibilities are not equivalent to those of an administrative manager.

A cybersecurity architect and a financial controller evaluate different dimensions of enterprise risk.

The objective is not to eliminate functional expertise.

It is to eliminate unnecessary fragmentation in how that expertise contributes to business outcomes.

This distinction suggests a different organizational design principle.

Preserve specialization where it creates expertise, accountability, or control. Reduce boundaries where they primarily create information-transfer overhead.

For example, a procurement process may still require legal review of unusual contractual terms.

But it may not require legal personnel to manually retrieve the same contract information already available through governed enterprise services.

Finance may still retain authority over payment controls.

But it may not need to receive manually assembled reports before evaluating a transaction that an integrated system can present with authoritative supporting information.

The value lies in preserving necessary judgment while reducing unnecessary coordination.

AI can support this transition.

Enterprise leadership must design it.

Management Layers Will Face Different Economic Pressures

One of the more controversial implications of AI is its potential effect on management hierarchy.

It would be premature to conclude that AI will universally flatten organizations.

Management layers serve multiple purposes.

They coordinate work.

They develop employees.

They resolve conflicts.

They allocate resources.

They maintain organizational accountability.

They provide continuity and leadership.

Some of these responsibilities may become more demanding as AI changes the enterprise.

Others may require less human effort.

The outcome will vary by industry, organizational scale, operating model, and management capability.

Nevertheless, the economic assumptions supporting traditional spans of control deserve reconsideration.

Suppose a manager supervises eight employees.

A substantial portion of the manager’s time may involve collecting status updates, reviewing routine work, preparing reports, and coordinating information.

If AI reduces the effort required for those activities, the manager may have more capacity to focus on coaching, complex decisions, and performance improvement.

That could support a broader span of control.

But the opposite may occur in some environments.

If AI agents introduce complex operational risks, unfamiliar workflows, or significant exception-handling responsibilities, management attention requirements may increase.

The appropriate span of control depends on the nature of the work, not simply the number of people or agents involved.

An enterprise should therefore avoid simplistic assumptions that AI will automatically eliminate middle management.

The more useful question is:

Which management responsibilities are becoming less expensive, and which are becoming more consequential?

Organizational design should follow that analysis.

The Emerging Distinction Between Information Management and Decision Management

As AI assumes more information-processing activities, management responsibilities may increasingly separate into two categories.

The first is information management.

This includes collecting reports, consolidating operational data, distributing updates, preparing routine analyses, and maintaining awareness of activities.

The second is decision management.

This includes determining priorities, evaluating tradeoffs, allocating resources, resolving conflicts, exercising authority, and accepting accountability for outcomes.

These categories overlap.

Managers often perform both simultaneously.

But AI makes the distinction strategically important.

An organization may no longer require several layers of manual information consolidation to support an executive decision.

It may still require those executives to exercise judgment about the consequences of the decision.

Consider an enterprise capital allocation process.

Historically, department managers prepare investment requests.

Finance consolidates the requests.

Business unit executives prioritize them.

A corporate committee evaluates competing proposals.

Senior leadership approves the final portfolio.

AI could accelerate financial modeling, scenario analysis, dependency identification, and comparison of investment alternatives.

That does not eliminate the need to decide which investments deserve funding.

It changes how much organizational effort must be devoted to preparing the decision.

The enterprise should not confuse the reduced cost of decision preparation with reduced responsibility for decision quality.

As information processing becomes less expensive, the comparative value of legitimate decision authority may increase.

That has implications for leadership development, managerial competencies, and organizational structure.

The managers who create the greatest value may increasingly be those who exercise sound judgment, develop people, and improve business outcomes—not those who control the movement of information.

From Reporting Hierarchies to Decision Architectures

A traditional organizational chart answers several familiar questions.

Who reports to whom?

Who manages which department?

Who supervises which employees?

Where does authority reside?

These questions remain necessary.

But an AI-enabled enterprise needs an additional architectural view.

It must understand how decisions occur.

For each consequential decision, the enterprise should know:

• What business outcome the decision supports.
• Which role possesses legitimate decision authority.
• Which information sources are authoritative.
• What alternatives and constraints must be evaluated.
• Which activities AI may perform.
• What conditions require human judgment or approval.
• How the decision is executed.
• What evidence demonstrates authorization and execution.
• How the resulting outcome is measured.

This creates a decision architecture that complements the organizational chart.

The organizational chart establishes reporting and management relationships.

The decision architecture establishes how authority is exercised within business processes.

The two should be consistent.

But they are not identical.

A manager may supervise employees who participate in several business processes without personally authorizing every decision those processes require.

An executive may retain accountability for an outcome even when authorized decisions are executed by AI agents.

A cross-functional process may depend on several distinct decision authorities.

The enterprise must represent these relationships explicitly.

A Decision-Centric Organizational Design Framework

Design Dimension	Traditional Emphasis	AI-Enabled Emphasis
Organizational structure	Departments and reporting relationships	Business outcomes and decision responsibilities
Information flow	Reports, meetings, and managerial escalation	Governed, context-aware information access
Management activity	Coordination and activity supervision	Judgment, exceptions, and outcome supervision
Decision authority	Roles, policies, and informal practices	Explicit, enforceable decision rights
Process execution	Human handoffs across functions	Integrated human-agent workflows
Performance measurement	Departmental output and utilization	End-to-end performance and value realization
Accountability	Reporting hierarchy	Named ownership linked to decisions and outcomes

This is not a proposal to eliminate the organizational chart.

It is a proposal to stop treating the organizational chart as a sufficient description of how the enterprise operates.

The future enterprise will need to understand not only who reports to whom, but who—or what—can make which decisions, using what information, under whose authority.

That distinction becomes increasingly important as AI agents move from preparing recommendations to executing authorized actions.

Organizational Design Must Follow Business Architecture

Enterprise architecture provides a useful foundation for this transformation.

Business architecture describes the capabilities, value streams, processes, information, and organizational relationships required to execute strategy.

Organizational structure is one element of that architecture.

It should not be treated as an immutable starting point.

Consider a business capability such as customer onboarding.

The enterprise can examine the capability independently of the departments currently responsible for its activities.

It can identify the required outcomes, decisions, information, controls, and resources.

It can then determine how people, AI systems, and enterprise applications should contribute.

Only after those relationships are understood should leadership determine the most appropriate organizational responsibilities.

This reverses a common implementation approach.

Traditionally, organizations often begin with existing departments and ask how AI can make each department more productive.

A business architecture approach begins with the enterprise capability and asks how it should operate.

The distinction is consequential.

The first approach tends to preserve existing boundaries.

The second creates an opportunity to redesign them.

For example, a customer onboarding capability may involve sales, legal, finance, IT, and operations.

The enterprise might establish a single accountable process owner while preserving specialized authority within each function.

AI could coordinate information retrieval, validation, workflow progression, and routine execution.

Human specialists would remain responsible for decisions requiring their expertise.

The organizational design would reflect the needs of the business capability rather than the historical sequence of departmental activities.

Business architecture should inform organizational design. Organizational history should not permanently constrain business architecture.

A Practical Example: Redesigning Enterprise Customer Onboarding

Consider a business-to-business technology provider onboarding enterprise customers.

The existing process involves six departments.

Sales collects customer requirements.

Legal negotiates contractual conditions.

Finance establishes billing arrangements.

Security reviews access and compliance requirements.

IT provisions the necessary systems.

Customer success coordinates implementation and ongoing service.

Each department has its own management structure, performance measures, and queue of work.

The process takes an average of 15 business days.

Now suppose the organization introduces AI capabilities.

Sales representatives use AI to prepare customer documentation.

Legal uses AI to summarize contracts.

Finance uses AI to validate billing information.

IT uses AI to prepare provisioning requests.

Each department reports meaningful productivity improvements.

But customer onboarding still takes 15 business days.

Why?

Because the organization improved individual activities without redesigning the coordination architecture.

Approvals still occur sequentially.

Information is still transferred between departments.

Exceptions still move through multiple management layers.

No single operational view connects the entire process.

The enterprise has accelerated tasks without removing the constraints governing the customer outcome.

The Alternative: An Outcome-Centered Operating Model

Now consider a redesigned process.

The organization establishes customer onboarding as an end-to-end business capability with a named accountable owner.

A governed workflow retrieves relevant information from authoritative systems.

AI prepares required documentation and identifies missing information.

Independent activities proceed concurrently where appropriate.

Legal reviews contractual exceptions rather than every routine provision.

Finance evaluates transactions requiring financial judgment.

Security enforces access and control requirements.

IT provisioning begins when the required authorizations are satisfied.

Customer success monitors the complete onboarding journey.

The organization retains the specialized responsibilities of its departments.

But the process no longer depends on every department manually coordinating its portion of the information flow.

This may reduce cycle time, rework, and customer friction.

More importantly, it changes the organizational conversation.

Instead of asking whether each department is operating efficiently, leadership asks whether the enterprise is delivering the customer outcome efficiently.

The difference is substantial.

AI creates the opportunity to organize coordination around the customer outcome rather than the departmental reporting structure.

The actual benefits would need to be validated through process performance and financial measurements, not assumed from the redesign alone.

The New Economics of Span of Control

One of the most important organizational design questions concerns managerial capacity.

Traditional spans of control reflect the time and attention required to supervise employees.

AI may change both.

A manager equipped with reliable operational intelligence may spend less time collecting status reports and more time addressing exceptions.

Automated workflow monitoring may reduce the effort required to track routine execution.

AI-generated analysis may help identify emerging performance problems.

These capabilities could allow some managers to supervise broader operations.

But managerial capacity is not determined solely by information availability.

Employees require coaching, professional development, conflict resolution, and leadership.

Complex decisions require judgment.

Organizational change requires communication and trust.

Agentic workflows introduce additional responsibilities involving authority, exception handling, and operational assurance.

Consequently, the appropriate managerial span of control should be evaluated through several dimensions.

Dimension	Organizational Design Question
Work complexity	How much professional judgment and supervision does the work require?
Information burden	How much management effort is devoted to collecting and interpreting information?
Decision consequences	What financial, operational, or human consequences arise from decisions?
Employee development	How much coaching, mentoring, and professional guidance is necessary?
AI-enabled execution	Which routine activities can be reliably automated or coordinated?
Exception workload	How frequently must managers intervene in unusual or consequential situations?
Accountability	What outcomes and risks remain under the manager’s responsibility?

This analysis is more useful than imposing a universal target for management-layer reduction.

An organization may discover that some managerial spans can expand.

Others may need to remain unchanged.

Some may require additional specialized oversight.

The objective is not organizational flattening for its own sake.

It is designing management capacity around the work that genuinely requires management.

AI Could Strengthen Middle Management—or Expose Its Weaknesses

Middle management is frequently discussed as a likely target for AI-driven restructuring.

That conclusion is too simplistic.

Some middle-management activities involve information consolidation and routine coordination.

Those activities may become substantially easier to perform with AI.

But effective middle managers also translate strategy into execution.

They develop employees.

They resolve competing priorities.

They identify operational risks.

They manage organizational change.

They exercise judgment under uncertainty.

These responsibilities may become more important as AI accelerates the pace of business activity.

The distinction is between managers who primarily transmit information and managers who improve organizational performance.

AI may reduce the value of the former activity.

It may increase the leverage of the latter.

Consider a manager who spends much of the week preparing reports for senior leadership.

AI could reduce that administrative burden.

The manager might then devote more time to improving processes, developing employees, and resolving operational problems.

That could increase the value of the position.

But if the organization continues evaluating the manager primarily on reporting activity, it may fail to recognize the opportunity.

This is why organizational redesign must include performance management.

The enterprise should not simply remove information-processing work from managers. It should redefine the value management is expected to create.

The Risk of Replacing Human Bottlenecks With Algorithmic Bottlenecks

AI can reduce information-processing delays.

But poorly designed AI systems can introduce new constraints.

An enterprise may centralize access to information through a single AI platform.

It may require every decision to pass through an automated evaluation service.

It may deploy agents that depend on unreliable data integrations.

It may create excessive exception queues.

It may require human approval of thousands of routine recommendations.

These arrangements can recreate the same coordination problems the technology was intended to solve.

In some cases, they may create new dependencies that are less visible than traditional organizational bottlenecks.

For example, an agent may coordinate procurement activities across several applications.

If the agent relies on a single outdated supplier record, its recommendations may become systematically unreliable.

If the agent cannot distinguish authoritative data from conflicting departmental records, it may repeatedly escalate decisions.

If every proposed transaction requires managerial approval regardless of risk, the process may accumulate a growing queue.

The enterprise has replaced manual coordination with automated coordination but has not necessarily improved the underlying process.

This demonstrates why organizational redesign cannot be separated from data architecture, decision authority, and operational controls.

AI-enabled information flow must be reliable, authorized, and appropriate to the business context.

Otherwise, the organization may simply relocate its bottlenecks.

The Human Dimension: Organizations Are Not Just Information Networks

There is a danger in viewing organizational structures exclusively through the economics of information.

Enterprises are also social systems.

Employees need leadership, trust, professional identity, development opportunities, and meaningful relationships.

Teams create shared knowledge and organizational resilience.

Managers help employees interpret priorities and navigate uncertainty.

Informal communication often reveals problems that formal reporting systems fail to capture.

AI can support some of these activities.

It does not eliminate their importance.

An organization that removes management layers solely because AI can produce reports faster may unintentionally weaken coaching, employee engagement, or operational judgment.

A department that eliminates established collaboration mechanisms may lose important institutional knowledge.

A business that automates every available coordination activity may reduce opportunities for employees to identify emerging problems through professional interaction.

These consequences should be considered during organizational redesign.

The objective is not to minimize human involvement indiscriminately.

It is to distinguish between human interaction that creates value and administrative coordination that exists primarily because information is difficult to move.

In healthcare, for example, AI may improve access to patient information and reduce administrative work.

But clinical supervision, professional collaboration, and patient relationships remain essential.

In engineering, AI may accelerate technical analysis.

But peer review, mentoring, and collective problem-solving continue to support quality and safety.

In enterprise leadership, AI may improve access to operational intelligence.

But executives remain responsible for judgment, organizational direction, and accountability.

The AI-enabled organization should reduce unnecessary information friction without reducing the human relationships that make effective organizations possible.

The Workforce Development Implications

If AI changes the information-processing responsibilities embedded in organizational structures, it will also change the competencies required of managers and employees.

Traditional management development frequently emphasizes planning, reporting, delegation, coordination, and performance evaluation.

Those capabilities remain relevant.

But AI-enabled management requires additional skills.

Managers must understand how AI-generated information is produced and where its limitations lie.

They must recognize when automated recommendations require independent validation.

They must understand delegated decision authority.

They must supervise workflows involving both people and software.

They must evaluate process outcomes rather than relying exclusively on activity measures.

They must recognize when automation creates new operational risks.

They must also become more effective at the responsibilities that remain distinctly human: leadership, judgment, coaching, conflict resolution, and organizational development.

For employees, the implications are equally significant.

Some roles may shift away from routine information processing toward exception handling, analysis, relationship management, and business improvement.

Other roles may require deeper technical or professional specialization.

New responsibilities may emerge around agent supervision, workflow design, data authority, and operational assurance.

These transitions will not occur automatically.

Organizations need deliberate workforce development strategies.

Training employees to use AI tools is necessary but insufficient.

They must also understand how their responsibilities and decision authority are changing.

AI literacy teaches employees how to use the technology. Organizational readiness prepares them to operate within the enterprise that technology is creating.

A Framework for Evaluating Organizational Redesign

Before changing reporting relationships or management structures, executives should evaluate how AI affects the work their organizations perform.

I recommend a six-stage assessment.

Stage 1: Map the Existing Information Flow

Identify how information moves through a business capability.

Document where it originates, who collects it, how it is interpreted, and which management layers receive it.

Pay particular attention to recurring reports, approval queues, departmental handoffs, and duplicated information processing.

Stage 2: Identify the Actual Decision Points

Separate activities that prepare information from decisions that require legitimate authority.

Determine which decisions are routine, which involve material consequences, and which require professional judgment.

This distinction prevents the organization from treating every reporting activity as a necessary management responsibility.

Stage 3: Evaluate AI-Enabled Information Capabilities

Determine which information-processing activities AI can reliably improve.

Assess retrieval, reconciliation, summarization, analysis, workflow coordination, and exception detection.

Evaluate data quality, authoritative sources, integration requirements, and technical limitations.

Stage 4: Redesign the End-to-End Workflow

Examine whether existing departmental sequences remain necessary.

Identify opportunities for parallel execution, reduced handoffs, improved coordination, and more direct access to authoritative information.

Preserve required controls and specialized judgment.

Stage 5: Reallocate Management Responsibilities

Determine how changes in information processing affect managerial workload, spans of control, decision authority, and accountability.

Redefine roles around the management value that remains necessary.

Do not assume that every reduction in administrative effort justifies eliminating a position.

Stage 6: Measure Organizational Outcomes

Evaluate whether the redesign improves business performance.

Relevant measures may include decision latency, process cycle time, management overhead, employee engagement, customer outcomes, operational risk, and financial performance.

The enterprise should also monitor unintended consequences.

A structure that reduces management expense but increases errors, employee turnover, or operational risk may not represent an improvement.

The objective is sustainable organizational effectiveness.

Measuring the Organizational Impact of AI

Traditional AI adoption metrics are inadequate for evaluating organizational redesign.

The number of employees using AI does not establish whether information moves more effectively.

The number of automated reports does not demonstrate improved decision quality.

The number of management layers removed does not establish increased organizational performance.

Executives need measures that connect information processing to decisions and business outcomes.

A useful scorecard might include:

Measurement Area	Illustrative Indicators
Information latency	Time between an operational event and availability of relevant information
Decision latency	Time between decision readiness and authorized action
Coordination overhead	Handoffs, duplicate reporting, and time spent reconciling information
Decision quality	Error rates, reversals, exceptions, and outcome effectiveness
Management effectiveness	Span of control, coaching capacity, and exception-resolution performance
Employee experience	Role clarity, engagement, workload, and professional development
Business performance	Cycle time, service quality, operating cost, and value realization

These measures should be evaluated together.

Reducing decision latency is not beneficial if decision quality deteriorates.

Increasing managerial spans of control is not beneficial if employees lose necessary support.

Reducing reporting activity is not beneficial if executives become less aware of material operational conditions.

Organizational redesign requires a balanced view of efficiency, effectiveness, and accountability.

The CIO’s Role in Organizational Redesign

The CIO is not traditionally responsible for designing the entire enterprise organizational chart.

That responsibility generally belongs to executive leadership in coordination with business functions and human resources.

But AI makes the CIO an increasingly important participant in organizational design.

The architecture of enterprise information influences how work can be coordinated.

Data integration determines whether functions share reliable information.

Identity and authorization services determine how access and actions are controlled.

Workflow orchestration influences how activities move across systems.

Agentic capabilities affect where work is executed.

Observability determines whether management can evaluate operational performance.

These architectural decisions create possibilities—and constraints—for organizational redesign.

A fragmented technology environment may require extensive manual coordination even when AI capabilities are available.

An integrated, governed architecture may allow the enterprise to redesign workflows around outcomes rather than application boundaries.

The CIO must therefore collaborate closely with business executives, enterprise architects, the CFO, and workforce leadership.

The goal is not to use technology architecture to dictate organizational structure.

It is to ensure that organizational decisions reflect what the enterprise’s technology and information capabilities can reliably support.

The CIO’s strategic contribution is to make new organizational possibilities technically feasible without allowing technology to determine business accountability.

The CEO’s Role: Redesigning the Enterprise, Not Just Deploying AI

Ultimately, organizational design is an executive leadership responsibility.

The CEO and executive team must determine how the enterprise should organize its capabilities to execute strategy.

AI changes the assumptions underlying those decisions.

If information can move across functions more efficiently, some coordination structures may become less necessary.

If AI agents can execute bounded activities, management responsibilities may shift toward supervision and exception handling.

If business processes can operate across traditional departmental boundaries, accountability may need to be defined around end-to-end outcomes.

These changes cannot be managed exclusively through departmental technology initiatives.

They require enterprise-level decisions.

The CEO must ensure that business leaders do not optimize their individual departments at the expense of overall performance.

The CFO must evaluate the economic implications.

Human resources must address workforce development and organizational consequences.

The CIO must establish enabling architecture.

Governance functions must preserve appropriate authority and control.

Together, these leaders must determine whether the existing organizational structure remains appropriate for the enterprise’s emerging operating model.

The objective should not be a predetermined reduction in management layers.

It should be an organization better suited to achieving its strategic objectives.

What Boards Should Ask About Organizational Readiness

Boards should expect management to explain how AI may affect organizational effectiveness.

That discussion should extend beyond workforce reduction forecasts and productivity estimates.

Relevant questions include:

1. Which organizational activities exist primarily to collect, consolidate, or distribute information?
2. Where do departmental boundaries create unnecessary delays or duplicated work?
3. Which decisions could be made more efficiently with reliable AI-enabled information?
4. Are decision rights and accountability sufficiently explicit to support new workflows?
5. How might AI change management responsibilities and appropriate spans of control?
6. What workforce development investments will be necessary?
7. How will leadership preserve professional judgment, employee support, and organizational resilience?
8. What measures will demonstrate that organizational changes improve business outcomes?

These questions do not presume that every organization should become flatter.

They require management to evaluate whether the enterprise’s structure remains aligned with its capabilities and strategic needs.

That is an appropriate board-level concern.

The Boardroom Question

If we were designing our enterprise today—with reliable, governed AI capabilities available from the beginning—would we create the same reporting relationships, departmental handoffs, management layers, and decision processes that we currently maintain?

The answer may be yes for some parts of the organization.

For others, the answer may be substantially different.

What matters is that leadership can explain the rationale.

Organizational structures should exist because they support effective execution, not merely because they reflect historical practice.

The Executive Imperative

Enterprise leaders should begin evaluating organizational structures through the lens of information flow and decision architecture.

This does not require immediate restructuring.

It requires understanding which organizational activities are changing and why.

Start with business capabilities where information delays, duplicated coordination, or fragmented decision-making materially constrain performance.

Map the existing information flow.

Identify the decision authorities.

Evaluate where AI can reduce coordination costs.

Redesign the workflow.

Then determine whether management responsibilities, reporting relationships, or departmental boundaries should change.

The sequence matters.

Organizations that begin with a target number of management layers risk making structural changes without understanding the work those layers perform.

Organizations that begin with business architecture can make more informed decisions.

They can preserve valuable expertise and leadership while reducing unnecessary information friction.

They can establish explicit decision authority.

They can create clearer accountability for end-to-end outcomes.

And they can prepare employees and managers for responsibilities that will continue evolving as AI capabilities mature.

The objective is not the smallest possible organization. It is the most effective organization for the work the enterprise must perform.

Conclusion: The Organizational Chart Must Evolve With the Economics of Information

The traditional organizational chart reflects an era when information moved primarily through people.

Departments collected and interpreted information.

Managers coordinated its movement.

Hierarchies established authority and provided mechanisms for resolving complexity.

Those structures helped enterprises scale.

They will continue to serve important purposes.

But artificial intelligence is changing the economics of information processing.

It can reduce the effort required to retrieve, consolidate, interpret, and distribute information.

It can support decisions across organizational boundaries.

AI agents can extend those capabilities into coordinated execution.

As these capabilities mature, some organizational structures may become less necessary, while others become more important.

Management may shift from supervising information flow toward exercising judgment and accountability.

Departments may retain specialized expertise while participating in more integrated workflows.

Decision authority may become more explicit.

Business capabilities may become more important units of organizational design.

The enterprise will still need leaders.

It will still need managers.

It will still need specialized functions.

But it should reconsider whether every existing reporting relationship and coordination mechanism remains justified.

The organizational chart was designed for human information flow. The intelligent enterprise must be designed for governed information, effective decisions, and accountable execution.

The companies that recognize this distinction will have an opportunity to redesign their organizations around what creates value.

Those that ignore it may continue operating with management structures optimized for constraints that technology has already begun to remove.

The future of organizational design will not be determined by how many people AI replaces.

It will be determined by how intelligently enterprises reorganize the relationship between information, decisions, people, and technology.

References

Autor, David H. “Why Are There Still So Many Jobs? The History and Future of Workplace Automation.” Journal of Economic Perspectives 29, no. 3 (2015): 3–30.

Brynjolfsson, Erik, and Lorin M. Hitt. “Beyond Computation: Information Technology, Organizational Transformation and Business Performance.” Journal of Economic Perspectives 14, no. 4 (2000): 23–48.

Coase, R. H. “The Nature of the Firm.” Economica 4, no. 16 (1937): 386–405.

Galbraith, Jay R. Designing Complex Organizations. Reading, MA: Addison-Wesley, 1973.

Malone, Thomas W. The Future of Work: How the New Order of Business Will Shape Your Organization, Your Management Style, and Your Life. Boston: Harvard Business School Press, 2004.

March, James G., and Herbert A. Simon. Organizations. New York: Wiley, 1958.

Mintzberg, Henry. The Structuring of Organizations: A Synthesis of the Research. Englewood Cliffs, NJ: Prentice-Hall, 1979.

Ross, Jeanne W., Peter Weill, and David C. Robertson. Enterprise Architecture as Strategy: Creating a Foundation for Business Execution. Boston: Harvard Business School Press, 2006.

Simon, Herbert A. Administrative Behavior: A Study of Decision-Making Processes in Administrative Organization. New York: Macmillan, 1947.

Williamson, Oliver E. The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting. New York: Free Press, 1985.

Coming Next

Article 14: The AI Skills Gap Is Not What Most Companies Think It Is

Enterprise discussions about the AI skills gap frequently focus on technical capabilities: prompt engineering, machine learning, model development, and the ability to operate AI tools.

Those skills matter.

But they do not fully address the workforce challenge facing organizations.

As AI changes how work is organized, employees and managers will need different combinations of business judgment, process understanding, information literacy, decision-making capability, and technological fluency.

The next article examines why the most consequential AI skills gap may not be the shortage of employees who understand artificial intelligence.

It may be the shortage of leaders and professionals who understand how to redesign business work around it.

The intelligent enterprise will require more than employees who know how to use AI. It will require people who know how to make AI-enabled organizations work.