Site icon Chief Magazine

Chief AI Leadership Pulse 2026

Chief Magazine
AI Leadership_ Business Struggles Unveiled (1)

What Business Leaders Are Really Struggling With as AI Moves From Experimentation to Measurable Value

October 2026 | Chief Magazine Research & Analysis

AI has become much easier to experiment with. The harder question for business leaders in 2026 is what happens after the experiment.

Companies have spent the past few years adding copilots, testing generative AI, automating individual tasks and running proofs of concept across departments. Employees are increasingly comfortable using AI for everyday work, and the technology itself is improving at a remarkable pace. Yet the financial impact at the enterprise level remains much less obvious.

That gap is becoming one of the central management questions of the AI era: How does a company move from having people use AI to actually changing the economics of the business?

The latest research from PwC and McKinsey suggests that this is where the real leadership challenge now sits. PwC’s 2026 Global CEO Survey found that only 12% of CEOs reported that AI had delivered both cost and revenue benefits, while 56% reported neither significant cost reduction nor revenue growth from AI. McKinsey’s 2026 State of AI survey similarly found that 80% of respondents said AI had improved their individual productivity, but only 37% attributed any positive EBIT impact to AI at the organizational level.

The message is not that AI is failing. It is that AI adoption and AI value are turning out to be two different problems.

The AI adoption problem is becoming a business problem

For much of the last few years, the question inside companies was relatively straightforward: Where can we use AI?

That question has produced plenty of answers.

Marketing teams can generate and personalize content. Sales teams can summaries customer interactions. Developers can use coding assistants. Customer-service teams can automate routine enquiries. Finance teams can extract information from documents and accelerate analysis. Executives can use AI to synthesize large amounts of information.

The technology is increasingly capable of doing these things.

But successful individual use does not automatically translate into business transformation.

McKinsey’s August 2026 survey illustrates the distinction clearly. Four in five respondents said AI had improved their individual productivity, yet only 37% said their organization had achieved any positive EBIT impact from AI. Only about 6% met McKinsey’s definition of an AI high performer—companies attributing at least 5% of EBIT to AI and describing its impact as significant.

That creates a new challenge for CEOs.

The question is no longer simply whether employees are using AI. It is whether the organization has changed enough for that usage to appear in revenue, margins, customer experience, cycle times, decision quality or other meaningful business outcomes.

1. The first struggle: proving where the value actually comes from

AI creates an awkward measurement problem.

An employee who uses an AI assistant may finish a report faster. A salesperson may prepare for a meeting more efficiently. A developer may produce code more quickly. A customer-service representative may handle more interactions.

Those are useful outcomes, but they don’t necessarily show up immediately in the company’s financial statements.

Was revenue higher because of AI? Were costs lower because of AI? Did the company avoid hiring additional employees? Did customer retention improve? Did the organization launch products faster? Did better decisions create measurable value?

These questions are much harder to answer.

PwC’s January 2026 Global CEO Survey found that 30% of CEOs reported increased revenue from AI, while 26% reported decreased costs. But only 12% reported both outcomes.

The implication for leadership is important: AI ROI cannot be reduced to counting licenses, prompts or hours supposedly saved.

Companies need to connect AI initiatives to business metrics that already matter.

If an AI project is intended to improve customer service, for example, the measurement framework might include resolution time, cost per interaction, customer satisfaction and retention—not simply the number of employees using an AI assistant.

That sounds obvious. In practice, it requires considerably more discipline than launching another pilot.

2. The second struggle: too many pilots, not enough redesign

One of the most persistent problems with enterprise AI is that companies often add AI on top of existing work rather than changing how the work itself is performed.

An organization might introduce a sales copilot without changing its sales process. It might add an AI assistant to customer service without redesigning escalation workflows. It might give developers coding agents while keeping every surrounding approval and testing process exactly as before.

The result can be higher individual productivity without much change to the underlying economics.

McKinsey’s 2026 research provides an important contrast. Nearly three-quarters of its AI high performers reported fundamentally redesigning workflows because of AI, compared with only about one-quarter of other respondents. High performers were also twice as likely to report that senior leaders demonstrated commitment to AI initiatives and that their organizations had defined processes for measuring impact.

PwC found something similar from a different angle. Its study of 1,217 senior executives found that the top 20% of companies captured 74% of AI-driven economic value. These companies were twice as likely to redesign workflows around AI rather than simply adding AI tools.

This may be one of the most important lessons emerging from the current AI cycle:

Buying AI is a technology decision. Redesigning work around AI is a management decision.

The second one is much harder.

3. The third struggle: legacy systems are still in the room

AI demonstrations tend to happen in clean environments.

Real businesses rarely operate that way.

Their data may sit across multiple systems. Processes may depend on spreadsheets, legacy applications, manual approvals and disconnected databases. Important information may not be consistently structured. Different departments may define the same customer, product or transaction differently.

AI can expose these weaknesses very quickly.

Recent research from BearingPoint, reported by Reuters in September 2026, found that fewer than one-third of surveyed companies had moved their AI initiatives beyond the pilot stage, despite nearly three-quarters reporting positive financial outcomes from AI projects. Regulatory issues and integration with legacy IT systems were among the principal obstacles to scaling.

McKinsey’s 2026 research also shows that organizational and technology constraints remain important barriers. Its State of Organizations research identified regulatory, ethical and legal concerns, organizational challenges, inadequate technology infrastructure and lack of clear strategy or leadership support among the barriers to adopting AI at scale.

This creates an uncomfortable reality for executives.

Sometimes the obstacle to an AI project isn’t the AI.

It is the company underneath it.

4. The fourth struggle: AI strategy is still too often separated from business strategy

There is a temptation to create an “AI strategy” as a separate corporate initiative.

But businesses do not ultimately need an AI strategy for its own sake.

They need a strategy for increasing revenue, improving margins, serving customers, developing products, operating more efficiently and competing more effectively.

AI is increasingly becoming one of the mechanisms through which those objectives can be achieved.

PwC’s research is particularly interesting here. The companies capturing the most value from AI are not simply using it to reduce costs. They are significantly more likely to use AI to pursue growth opportunities and reinvent their business models.

That distinction matters.

If the only question leadership asks is:

“How many employees can AI make more productive?”

the organization may find efficiency gains.

If the question becomes:

“What could our business do differently if intelligence became dramatically cheaper and more scalable?”

the strategic possibilities become much larger.

That might mean personalized services that were previously uneconomic, faster product development, new customer segments or entirely different operating models.

McKinsey’s research similarly finds that high-performing organizations pursue growth and innovation alongside efficiency.

5. The fifth struggle: deciding what should actually be automated

The arrival of AI agents makes this question even more important.

Traditional automation generally required organizations to define relatively explicit rules.

AI makes it possible to automate parts of processes that involve interpretation, generation, classification and decision-making.

That creates opportunity—but also a management challenge.

Not every task should be automated simply because it can be.

A useful framework for leaders is to ask:

Is the work repetitive?

Is the outcome measurable?

Is there enough volume for automation to create meaningful economics?

What happens when the AI gets something wrong?

Does the process require human judgment or accountability?

Can the organization monitor the system once it is operating at scale?

McKinsey’s recent research on agentic workflows makes a similar point: technically impressive AI does not automatically create attractive economics. The strongest opportunities tend to occur where there is sufficient scale, repeatability and automation potential for improvements to compound.

For CEOs, this means the AI roadmap increasingly needs to look like an operating-model roadmap, not a shopping list of AI products.

6. The sixth struggle: leadership has to change with the technology

There is a tendency to treat AI transformation as something that can be delegated to the CIO, CTO or a dedicated AI team.

Those teams are essential.

But the broader organizational consequences cannot be delegated quite so easily.

Who decides which processes should change?

Who decides which decisions can be automated?

Who redesigns roles?

Who determines what humans should continue to own?

Who accepts responsibility when an AI-enabled process fails?

And who decides whether an AI investment is creating enough value to continue?

Those are leadership questions.

McKinsey’s September 2026 analysis makes this point directly, arguing that CEOs have responsibilities they cannot simply hand off, including raising the organization’s ambition, rearchitecting how it works and resetting organizational culture.

The technology may be new.

The management problem is not.

What Business Leaders Should Be Asking Now

The next phase of AI adoption may therefore require fewer questions about which AI tool to buy and more questions about which part of the business should change.

A CEO considering an AI initiative could start with six questions:

1. What business outcome are we trying to improve?

Revenue, margin, speed, quality, customer experience, employee capacity or something else?

2. What process needs to change?

If the workflow remains unchanged, why should the economics change?

3. How will we measure the result?

Define the baseline before implementing the technology.

4. What organizational capability is missing?

Data? Skills? Governance? Technology integration? Change management?

5. Where does human judgment still matter?

Automation should not automatically mean removing people from the process.

6. What happens if this works?

This is perhaps the least discussed question.

A successful pilot can create its own problem if the organization isn’t prepared to scale it.

The AI Gap Is Becoming an Organizational Gap

The most interesting thing about the current AI cycle may not be the widening capability of the technology.

It may be the widening difference between organizations that can absorb that capability and organizations that cannot.

PwC’s research found that the top 20% of companies captured 74% of AI-driven economic value. McKinsey’s research similarly shows a small group of high performers separating themselves through workflow redesign, leadership commitment, measurement and broader transformation.

That doesn’t mean every company needs to transform at the same speed or spend the same amount.

It does suggest that simply having access to the latest models will not create a durable advantage.

The models are increasingly available to everyone.

The harder-to-copy asset is the organization that knows where to apply them, how to redesign work around them, how to measure the result and how to scale what works.

The Question CEOs Should Be Asking

The AI conversation has moved on.

A few years ago, the question was:

“Should we adopt AI?”

Then it became:

“Where can we use AI?”

Now a more consequential question is emerging:

“What should we change about the way our company works because AI exists?”

That is a much harder question.

It involves technology, but it also involves processes, people, incentives, governance, data, capital allocation and leadership.

The organizations that answer it well may not necessarily be the ones using the most AI.

They may be the ones that understand which parts of their business need to change—and have the discipline to change them.

The Chief View

The biggest AI leadership challenge in 2026 isn’t access to AI. It is converting access into organizational change.

The data increasingly points in the same direction: individual productivity gains are real, but enterprise-level financial impact remains much less widespread. The companies generating stronger returns are more likely to redesign workflows, pursue growth opportunities alongside efficiency, establish measurement systems and make AI part of how the organization actually operates.

For business leaders, that changes the job.

AI doesn’t need another executive sponsor who approves pilots.

It needs leaders willing to ask what the company should stop doing, start doing and do differently because intelligence is becoming more accessible.

That is where experimentation becomes transformation.

Sources & further reading

Exit mobile version