Future:Fit Leadership — Blog 9

The Augmented Operating Model

Most organizations are introducing AI into operating models designed before intelligent agents existed.

They add copilots to existing roles.

They automate individual tasks.

They create an AI centre of excellence.

Then they wait for transformation.

What they often get is scattered productivity: pockets of faster work surrounded by the same approvals, handoffs, functional boundaries and outdated management practices.

The technology changes.

The organization does not.

To capture the real value of AI, leaders must move beyond adding tools and begin redesigning how work flows through the enterprise.

Do not start with the organization chart

When leaders hear “organization design,” they often think about structure.

Who reports to whom?
Which roles should move?
How many layers are needed?
Where should AI capability sit?

These are important questions—but they come too early.

An organization chart shows the formal hierarchy.

It rarely shows how value is actually created.

It does not reveal the repeated handoffs, duplicate reviews, hidden workarounds, inaccessible data or decisions that travel up three layers before moving sideways into another function.

AI will not fix these problems automatically.

It may simply help them happen faster.

The right starting point is not structure.

It is work.

The unit of transformation is the workflow

Most early AI initiatives focus on isolated tasks: summarize this document, draft that email, analyze these figures.

But work is not a collection of independent tasks.

It is a connected flow.

Research highlighted by MIT Sloan argues that AI’s greatest value may emerge from redesigning entire workflows—how tasks are sequenced, grouped and handed between humans and machines—rather than optimizing individual activities.

This matters because every handoff creates friction.

If AI completes one step in seconds but the output then waits three days for approval, the system has not become meaningfully faster.

If an employee saves five hours preparing a report that nobody uses, productivity has increased while value remains unchanged.

If an agent generates customer insight but the frontline cannot act on it, intelligence has been created without agency.

The augmented operating model connects intelligence to action.

Design work before roles—and roles before structure

Future-fit organization design follows a different sequence:

Value → Work → Human–AI allocation → Decisions → Roles → Structure

1. Value: What outcome are we creating?

Begin with the customer, citizen, employee or stakeholder.

What outcome matters to them? What problem are we solving? What would better, faster or more personal value look like?

This prevents teams from automating work simply because it already exists.

Some activities should be improved.

Some should be combined.

Some should disappear.

2. Work: What must happen to create that value?

Map the end-to-end workflow across functional boundaries.

Where does work begin? What information is required? Where does it stop, wait or return for correction? Which steps create value, manage necessary risk or merely preserve habit?

Do not map the process as leaders imagine it.

Map the work as people actually perform it.

3. Allocation: What should humans and AI each do?

Examine the work at task and workflow level.

AI is often well suited to:

• Searching and synthesizing information
• Detecting patterns
• Generating drafts and alternatives
• Monitoring routine conditions
• Coordinating repeatable workflows
• Operating at scale and speed

Humans remain essential for:

• Defining purpose
• Navigating ambiguity and exceptions
• Exercising ethical judgment
• Building trust
• Creating meaning
• Managing conflict
• Owning consequential decisions

The goal is not maximum automation.

It is the best combination of machine capability and human contribution.

4. Decisions: Where should authority sit?

AI changes the speed at which information can become available.

But if every decision still moves upwards through the hierarchy, the operating model remains slow.

Ask:

• Which decisions can move closer to the work?
• What can an AI agent recommend?
• What can it execute within defined boundaries?
• Which decisions require human approval?
• Which decisions must remain entirely human?
• When should an exception be escalated?

Decision rights are the nervous system of the operating model.

If they remain unclear, speed creates confusion rather than agility.

5. Roles: How does human contribution change?

Do not assume that automating tasks simply reduces the number of people required.

A role is a bundle of activities, relationships and accountabilities.

When AI takes on part of that bundle, the remaining role must be redesigned deliberately.

A manager may spend less time assembling reports and more time interpreting evidence, coaching people and acting on exceptions.

An analyst may move from producing information to testing assumptions and influencing decisions.

A customer-service professional may handle fewer routine enquiries but more complex and emotionally demanding situations.

Job quality does not improve automatically when routine work disappears.

Leaders must decide what better human work will replace it.

6. Structure: What form best supports the new work?

Only now should leaders consider reporting lines, team boundaries, spans, layers and centralization.

The right structure may include:

• Smaller cross-functional teams organized around outcomes
• Shared AI, data and governance platforms
• Federated expertise embedded within business units
• Fewer coordination layers
• New roles overseeing human–agent workflows
• Communities that spread learning across the enterprise

Structure should support the flow of value.

The flow of value should not be forced to navigate an inherited structure.

The six elements of an augmented operating model

A coherent operating model aligns six elements:

Workflows: How value moves from need to outcome.

Decision rights: Who—or what—can decide and act.

Human and agent roles: How people and AI collaborate.

Data and platforms: What shared intelligence makes the work possible.

Governance: How risk, ethics, access and accountability are managed.

Measures: How the organization knows the new model is creating better outcomes.

Changing only one element creates tension.

An autonomous agent with no revised decision rights becomes an expensive assistant.

A redesigned workflow without accessible data becomes a diagram.

New roles without new performance measures are eventually pulled back towards old behaviour.

The elements must move together.

Centralize the foundations, distribute the innovation

Organizations often debate whether AI should be centralized or decentralized.

The answer is usually both—but for different purposes.

Centralize what must be trusted and reused:

• Platforms
• Security standards
• Data foundations
• Responsible-use principles
• Vendor controls
• Shared capabilities

Distribute what requires local knowledge:

• Use-case discovery
• Workflow redesign
• Experimentation
• Customer insight
• Adoption and learning

This creates freedom within a framework.

Teams can innovate without rebuilding the foundations or inventing governance each time.

The operating model must become a learning system

No organization can design the perfect human–AI model in advance.

Technology will change. Employee capability will grow. New risks and possibilities will emerge.

The operating model must therefore learn from its own work.

The Microsoft 2026 Work Trend Index describes leading organizations as learning systems—continually deciding what humans and agents should do and adapting how work is structured around them.

This requires short feedback loops.

Where did AI improve the workflow?
Where did it create a new bottleneck?
When did human intervention add value?
Which skills are becoming more important?
What should be standardized, scaled or stopped?

Organization design is no longer a periodic restructuring exercise.

It becomes a continuous leadership capability.

Your Future:Fit challenge: Choose one important customer or employee outcome. Map the workflow that produces it, then follow the sequence: value, work, human–AI allocation, decisions, roles and structure. Where is your current operating model preventing new intelligence from becoming better action?

The future-fit organization will not be defined by how much AI it owns.

It will be defined by how intelligently it organizes people, technology and work around value.

#FutureFitLeadership #AugmentedLeadership #OperatingModel #OrganizationDesign #FutureOfWork