By Vanguard Enterprise Intelligence Unit with the work of Rita McGrath, Amy Edmondson, Erik Brynjolfsson, Ethan Mollick, and Lynda Gratton.

Artificial intelligence has moved from a technology conversation to a strategy conversation. For many organizations, the question is no longer whether AI matters. It is whether the organization can absorb, apply, and adapt around AI fast enough to turn investment into competitive gain.

That distinction matters. Many companies are spending heavily on AI tools, pilots, platforms, consultants, data infrastructure, and internal experimentation. Yet the business impact often remains uneven. Some organizations achieve measurable gains in productivity, speed, personalization, risk detection, customer service, or decision quality. Others accumulate pilots without changing how work gets done.

The gap is rarely explained by technology alone. It is usually explained by organizational fitness.

In the AI era, “change fitness” is the capability to adapt repeatedly without losing strategic control. It is the ability of an organization to sense market shifts, redesign workflows, integrate human and machine judgment, update operating routines, and make decisions at the speed of changing evidence. It does not mean constant disruption. It means disciplined adaptation.

This is becoming one of the defining strategic capabilities of modern management. AI does not simply automate tasks. It changes the tempo of competition. It accelerates analysis, compresses planning cycles, expands the number of strategic options, and exposes weaknesses in decision systems that were designed for slower markets.

Organizations that treat AI as a tool purchase will usually underperform. Organizations that treat AI as an operating-model shift have a better chance of converting experimentation into advantage.

The Problem with AI Without Organizational Change

The early phase of enterprise AI adoption has produced a familiar pattern. A company identifies several promising use cases. Teams launch pilots. Executives communicate enthusiasm. Technology groups select platforms. Employees experiment with generative tools. Some productivity gains appear in isolated functions.

Then the momentum slows.

The pilot does not scale. The business case is unclear. Employees are uncertain when to trust AI outputs. Managers do not know how to measure improvement. Data quality problems surface. Legal and compliance teams raise concerns. The technology group becomes a bottleneck. Business units return to familiar processes. The organization has activity, but not transformation.

This pattern reflects a fundamental misunderstanding. AI value does not come from inserting a model into an existing organization and expecting performance to improve automatically. AI creates value when work is redesigned around new capabilities.

For example, a customer service team may use generative AI to draft responses faster. That may reduce handling time. But the larger opportunity may involve redesigning knowledge management, escalation rules, agent training, customer segmentation, quality review, and feedback loops. The tool improves a task. The operating model improves the system.

The same applies to strategy. AI can improve forecasting, scenario analysis, pricing, product development, fraud detection, procurement, marketing, and workforce planning. But if decision rights, incentives, data flows, and managerial routines remain unchanged, the organization will capture only a fraction of the potential value.

AI adoption is therefore not primarily a technical challenge. It is a change discipline.

From Five-Year Plans to Dynamic Strategy

The traditional strategic-planning model assumed a relatively stable planning horizon. Leaders assessed markets, set priorities, allocated capital, and built multi-year plans. Annual reviews adjusted the plan, but the core logic remained intact.

That model is increasingly strained. Competitive conditions now shift through technology adoption, customer behavior, platform dynamics, regulatory changes, capital-market pressure, supply-chain volatility, and labor-market transformation. AI intensifies this volatility because it changes what organizations can know, how quickly they can act, and how fast competitors can imitate or improve.

This does not mean long-term strategy is obsolete. It means long-term strategy must be managed through shorter learning cycles.

A dynamic strategy has three characteristics. First, it defines a clear strategic direction. Second, it converts that direction into near-term experiments, operating changes, and measurable capability-building. Third, it updates decisions based on real evidence.

The mistake is to confuse agility with improvisation. Organizations still need strategic priorities. They still need resource discipline. They still need executive alignment. But they also need planning systems that can incorporate new information quickly.

In this environment, a five-year plan should function less like a fixed map and more like a strategic thesis. It should define where the company intends to compete, what capabilities it must build, what assumptions must be tested, and what signals would require adjustment.

AI makes this possible, but only if the organization has the fitness to respond.

What Change Fitness Looks Like

Change fitness is not a cultural slogan. It is an operating capability. It can be observed in how an organization makes decisions, reallocates resources, redesigns roles, and learns from implementation.

A change-fit organization does four things well.

First, it senses change early. It does not rely only on quarterly reports or annual strategy reviews. It monitors customer behavior, competitor moves, employee adoption patterns, operational friction, and technology performance in near real time. AI can support this by detecting patterns, surfacing anomalies, and improving forecasting. But leaders must decide which signals matter.

Second, it translates signals into decisions. Many companies collect data faster than they can act on it. Change fitness requires decision pathways that are clear, timely, and accountable. When a signal appears, leaders need to know who owns the decision, what evidence is required, and how quickly resources can move.

Third, it redesigns work instead of layering tools on top of old processes. AI should not simply make outdated workflows faster. It should prompt leaders to ask which tasks should be automated, which should be augmented, which should remain human-led, and which should be eliminated altogether.

Fourth, it builds trust through disciplined adoption. Employees need clarity on how AI will be used, what risks must be managed, how performance will be measured, and where human judgment remains essential. Without trust, adoption becomes inconsistent. With blind trust, risk increases. Change fitness requires informed confidence.

Human-AI Teams as the New Unit of Performance

One of the most important shifts in the AI era is the emergence of human-AI teams. Competitive advantage will increasingly depend not only on individual talent or machine capability, but on the design of collaboration between people and intelligent systems.

In many organizations, the current model is informal. Employees use AI tools as personal assistants. They draft emails, summarize documents, generate ideas, write code, analyze data, or prepare presentations. This can be useful, but it is not enough. The larger strategic question is how human-AI collaboration should be built into the operating model.

Leaders should distinguish among three modes of human-AI work.

The first mode is automation. AI performs a task with limited human involvement. This is appropriate for repetitive, rule-based, or high-volume work where the risk is manageable and the output can be monitored.

The second mode is augmentation. AI improves human performance by expanding information access, generating options, identifying patterns, or accelerating analysis. The human remains responsible for judgment, context, and final decision-making.

The third mode is orchestration. Humans and AI systems interact across a workflow, with tasks moving between people, models, systems, and governance checkpoints. This is where the greatest organizational value often emerges, but it requires more deliberate design.

Human-AI teams should be designed around the work, not around the tool. Leaders should begin by identifying the decision or process that matters: approving credit, resolving customer issues, managing inventory, developing products, detecting fraud, designing campaigns, or forecasting demand. Then they should define where AI improves speed, where it improves quality, where it introduces risk, and where human oversight is non-negotiable.

The best organizations will not simply give employees access to AI. They will redesign team performance around AI-enabled workflows.

Sequencing Predictive and Generative AI

Many companies approach AI as if all AI investments belong in the same category. This creates confusion. Predictive AI and generative AI often serve different strategic purposes and require different adoption models.

Predictive AI is strongest where the organization needs to forecast, classify, detect, recommend, or optimize based on structured patterns. It can help predict churn, detect fraud, forecast demand, score leads, optimize pricing, identify maintenance risks, or improve logistics.

Generative AI is strongest where the organization needs to create, summarize, synthesize, explain, draft, simulate, or interact through language, code, images, or other content formats. It can help write proposals, generate product concepts, summarize research, support customer interactions, assist software development, or improve knowledge access.

The sequencing question is not which form is better. It is which form fits the strategic problem.

For some organizations, predictive AI should come first because the highest-value opportunities depend on better forecasting, risk detection, or operational optimization. For others, generative AI may create immediate value by reducing knowledge friction, improving content workflows, accelerating service responses, or supporting expert work.

A practical sequencing model should consider four factors: value, readiness, risk, and learning.

Value asks whether the use case is tied to a real business outcome. Readiness asks whether the organization has the data, workflow clarity, talent, and systems required to implement it. Risk asks what could go wrong if the AI output is inaccurate, biased, insecure, or misused. Learning asks whether the initiative builds reusable capability for future AI adoption.

This last point is important. The best AI projects do more than solve one problem. They build organizational muscle. They improve data discipline, governance, employee fluency, workflow redesign, and managerial confidence. Poorly chosen projects consume attention without increasing fitness.

Turning Skepticism into Strategic Momentum

Employee skepticism is often treated as resistance. That is too simplistic. Skepticism can be a useful signal. It may reveal unclear strategy, weak training, poor tool selection, fear of job displacement, or concern about quality and accountability.

Leaders should not try to overcome skepticism with broad enthusiasm. They should convert skepticism into structured learning.

The first step is to clarify the purpose of AI adoption. Employees need to understand whether the goal is cost reduction, quality improvement, speed, customer experience, risk management, innovation, or capacity expansion. Vague claims about transformation do not create trust.

The second step is to involve employees in workflow redesign. Frontline workers often understand process friction better than executives or technology teams. If they are excluded, AI adoption may solve the wrong problem.

The third step is to create safe experimentation boundaries. Employees should know what tools are approved, what data can be used, what outputs require review, and what use cases are off limits. Clear rules increase adoption because they reduce uncertainty.

The fourth step is to reward learning, not just success. Early AI adoption will involve mistakes. If employees believe every error will be punished, they will avoid experimentation or use tools quietly. Leaders should distinguish between reckless use and responsible learning.

The fifth step is to show evidence. Skepticism declines when people see practical improvements in their own work. The most persuasive argument for AI adoption is not executive messaging. It is a better workflow.

A Practical Model for Building Change Fitness

Leaders can build change fitness through a five-part operating model.

The first element is strategic intent. The organization must define what AI is expected to improve. Without strategic intent, AI adoption becomes scattered. Leaders should identify a small number of priority domains where AI can strengthen competitive position.

The second element is workflow redesign. Every significant AI initiative should be linked to a process map. What work is being changed? Who currently does it? What decisions are involved? What systems are touched? Where does AI enter? What human review is required?

The third element is capability building. Employees need training that is specific to their roles. General AI awareness is useful, but it does not substitute for functional fluency. Sales teams, finance teams, legal teams, operations teams, and product teams need different examples, safeguards, and performance measures.

The fourth element is governance. AI governance should not only prevent harm. It should enable responsible speed. Clear policies, approval pathways, model evaluation standards, data rules, and escalation protocols help teams move faster because they know the boundaries.

The fifth element is feedback infrastructure. Leaders need mechanisms to learn from adoption. Which use cases are creating value? Which are stalling? Where are employees confused? Where are customers affected? Which tools are being used outside formal channels? What risks are emerging? Change fitness depends on this feedback becoming visible and actionable.

Together, these elements convert AI from a series of experiments into an organizational capability.

The Leadership Discipline Required

AI places new demands on executives. Leaders do not need to become machine-learning engineers, but they do need to understand how AI changes the work of strategy.

They must ask better questions. What decision are we improving? What behavior must change? What capability will compound? What data do we need? What risks are acceptable? What governance is slowing us down unnecessarily? What human judgment must remain central? What will we stop doing if this works?

They must also model disciplined adaptation. If executives announce AI as a priority but keep legacy planning, budgeting, and performance systems unchanged, the organization will receive conflicting signals. People follow the system more than the speech.

The leadership challenge is to create enough urgency to move, enough structure to prevent chaos, and enough trust to sustain adoption.

Strategy as a Living System

The AI era requires organizations to treat strategy as a living system. The strategic direction should be clear, but the execution model must learn continuously. Market signals should feed planning. Employee experience should inform workflow redesign. AI outputs should improve decisions, but not replace accountability. Governance should protect the organization while enabling speed.

This is the practical meaning of change fitness.

Organizations with change fitness will not get every AI decision right. They will still face failed pilots, technical limits, cultural resistance, and uncertain returns. But they will learn faster. They will reallocate sooner. They will redesign work more intelligently. They will build trust through use rather than rhetoric.

Organizations without change fitness will continue to confuse AI activity with AI advantage. They will buy tools, launch pilots, and issue transformation language while the operating model remains mostly unchanged.

The difference will become more visible as AI becomes more embedded in competition. The winners will not simply be the companies with the most advanced models. They will be the companies that can organize people, processes, data, governance, and strategy around continuous adaptation.

In the AI era, strategic advantage will belong to organizations that can change deliberately, repeatedly, and faster than the market can punish delay.