The Sleepless Flight to New York

John Mercer should have been asleep.

The cabin lights were low. The hum of the engines was steady. Somewhere outside the small oval window, the Atlantic was an endless black sheet below the wing.

British Airways Flight BA117 had left Heathrow on time and was now somewhere over the ocean, suspended between London and New York, between one boardroom and another.

The aircraft was quiet in the way business-class cabins become quiet after dinner. Not silent, exactly. There was the soft clink of glassware from the galley. The occasional cough. The low murmur of a flight attendant asking someone if they wanted tea, coffee, or another glass of wine. A man two rows ahead was already flat under a blanket, mouth slightly open, sleeping with the confidence of someone who had either solved all his problems or learned to ignore them.

John envied him.

His own seat had been converted into a bed. A real bed, almost. White pillow. Soft blanket. Noise-canceling headphones. Perfect cabin temperature. A dinner he had barely touched but could not complain about. A glass of Bordeaux he had accepted out of habit and then forgotten beside his laptop.

Everything had been designed to help him rest.

None of it was working.

He looked at the screen in front of him.

Time to destination: 5 hours 42 minutes.

He had a meeting at 9:00 a.m. in Manhattan.

Not a casual meeting. Not one of those relationship meetings where people shake hands, exchange harmless optimism, and promise to “circle back.”

This one mattered.

The board wanted a plan.

The company had missed two product milestones. Their European operations team was overloaded. Their US customers were getting louder. The CFO had frozen new hiring three weeks earlier. The CTO had warned that engineering capacity was already stretched beyond reason. The COO had said, quite calmly, that “something would break” if they continued like this.

And John, as CEO, was expected to land in New York, look composed, and explain how they would move faster without spending more.

He closed his eyes.

That was a mistake.

The moment the cabin disappeared, the numbers arrived.

Twenty-seven open roles.

Six critical projects behind schedule.

Three enterprise customers waiting on integrations.

One security audit slipping into next quarter.

Two teams in London and Warsaw working late too often.

A New York sales team promising things delivery had not confirmed.

A finance team asking why vendor spend had increased while throughput had not.

An investor who had written, in a polite email that was somehow worse than an angry one:

“We need to see more leverage in the operating model.”

John opened his eyes again.

The flight attendant passed quietly through the aisle.

“Can I get you anything else, Mr. Mercer?”

He smiled automatically.

“No, thank you. I’m fine.”

It was a ridiculous answer.

He was not fine.

But business class had its own theater. People said they were fine. They smiled. They adjusted their blankets. They opened laptops at midnight. They pretended that being able to recline fully somehow meant they were not carrying a small company’s worth of anxiety across the ocean.

John turned back to his laptop.

The document on the screen was titled:

Q3 Operating Plan — Revised

He had changed the title four times.

The first version had been called Growth Plan.

That now felt dishonest.

The second had been Efficiency Plan.

Too cold.

The third had been AI Productivity Plan.

Too trendy.

Now it was Operating Plan — Revised, which sounded harmless enough to survive legal review, board scrutiny, and emotional exhaustion.

He scrolled to the section labeled Execution Capacity.

It was the section everyone cared about but nobody wanted to name directly.

The company needed more output.

Not more slides. Not more strategy. Not more alignment. Output.

Customers needed onboarding work finished. Product needed integrations shipped. Operations needed automation. Finance needed reporting rebuilt. Sales needed technical support. Compliance needed documentation. Customer success needed better tooling. The data team needed help cleaning up years of messy internal systems before anyone could talk seriously about AI.

The work was everywhere.

The capacity was not.

For months, every leadership meeting had circled the same drain.

“We need to hire.”

“We need to prioritize.”

“We need to push the vendors.”

“We need to automate.”

“We need to use AI.”

“We need to offshore more.”

“We need to bring more in-house.”

“We need to stop saying yes to everything.”

All of it was partly true.

That was the problem with executive conversations. The bad ideas were rarely completely wrong. They were just incomplete enough to be dangerous.

Hiring would help, eventually. But eventually was too late.

Prioritization was necessary. But prioritization did not make the unchosen work disappear; it only moved the disappointment to a different calendar month.

Vendors could help. But the current vendors behaved exactly like vendors. They waited for tickets, billed for activity, escalated ambiguity, and produced weekly reports that made delay sound professionally managed.

AI was powerful. Everyone agreed on that. The company had licenses, pilots, demos, internal champions, and at least one Slack channel where people posted breathless examples of what an AI agent had done in twenty-three seconds.

And yet, somehow, the roadmap was still late.

John stared at the words on the screen.

Increase productivity through AI-enabled delivery.

It sounded right.

It also sounded like something a person writes when he does not yet know how the work will actually get done.

He deleted the sentence.

Then he stared at the blank space it left behind.

Across the aisle, a woman in a navy blazer was reading a printed document with a pen in hand. Not a tablet. Not a laptop. Actual paper. She had the calm focus of a person reviewing something that mattered.

John noticed the logo at the top of the page. A private equity firm he recognized.

She caught him looking and gave the small smile strangers give on night flights. Not an invitation. Not rejection. Just acknowledgment.

“Couldn’t sleep either?” she asked.

John laughed softly.

“Clearly I’m not hiding it well.”

“No one on this flight is hiding it well,” she said. “Some are just better at lying down.”

He smiled for real this time.

She introduced herself as Elena Voss. Operating partner. Based between London and Boston. She worked with portfolio companies that had reached the stage where growth had become complicated and every answer had a cost attached.

John told her, carefully, that he ran a software company expanding across the US and Europe.

“Let me guess,” she said. “You need to grow revenue, reduce cost, improve delivery, use AI, avoid hiring too aggressively, keep the team motivated, and tell the board the model is scalable.”

John looked at her.

“That was uncomfortable.”

“It usually is.”

He closed his laptop halfway.

“Is there a script they hand out to all boards now?”

“No,” Elena said. “It’s just the same math everywhere.”

She placed her pen inside the document and closed it.

“Companies grew for years by adding people. More work meant more headcount. More customers meant more teams. More complexity meant more managers. More markets meant more offices. It all made sense until money got expensive, AI arrived, and everyone started asking why output does not scale as cleanly as payroll.”

John leaned back.

That was exactly it.

Payroll had scaled beautifully.

Output had not.

Every new department created new dependencies. Every new manager created new meetings. Every new tool created new dashboards. Every new vendor created new coordination work. Every new hire needed context, onboarding, permissions, priorities, and someone already overloaded to help them become useful.

The company had not become lazy.

It had become heavy.

That was harder to admit.

Lazy teams can be pushed. Heavy systems have to be redesigned.

Elena looked toward the dark window.

“I see the same pattern in almost every company now,” she said. “The board wants AI productivity. The executive team wants faster delivery. Finance wants lower fixed cost. Employees want less chaos. Customers want commitments kept. But the operating model still assumes work belongs inside fixed teams, fixed departments, fixed budgets, and fixed job descriptions.”

John said nothing.

She continued.

“So everyone tries to squeeze a new world through an old structure.”

There it was.

A sentence simple enough to be annoying.

A new world through an old structure.

John thought about his leadership team.

His CTO was brilliant and tired.

His COO was practical and increasingly blunt.

His CFO had become the adult in every room, which meant people had started avoiding her until they needed approval.

His HR leader was still talking about employer brand while everyone else was quietly asking whether they could afford the people already on the plan.

His product leaders wanted more autonomy.

His engineering leaders wanted fewer interruptions.

His customers wanted dates.

His board wanted leverage.

And John wanted, more than anything at that moment, a model that did not require him to choose between growth and weight.

The flight map now showed the aircraft halfway across the Atlantic.

Halfway.

That was the strange cruelty of night flights. You could be thousands of miles from where you started and still nowhere near where you needed to be.

Elena reopened her document.

“What are you really trying to solve?” she asked.

John almost answered automatically.

“Capacity.”

But he stopped himself.

That was the word they used internally. Capacity. A safe word. A board-friendly word. A word with spreadsheets attached to it.

But it was not the full truth.

“We have work that must get done,” he said slowly. “But every traditional way of getting it done creates another problem.”

Elena nodded.

“Good. That is more honest.”

He continued.

“If we hire, it takes months and increases fixed cost. If we outsource, we lose control or spend too much time managing the vendor. If we push internally, we burn people out. If we delay, we lose customer trust. If we buy software, we still need someone to implement and own the change. If we say AI will solve it, we sound modern but nothing actually lands unless someone redesigns the work around it.”

For the first time all evening, the problem sounded clear.

Not solved.

Clear.

And clarity, at 38,000 feet, felt like oxygen.

Elena tapped her pen once against the paper.

“Then your problem is not capacity.”

John frowned.

“It isn’t?”

“No. Your problem is execution architecture.”

He almost laughed.

“That sounds like something consultants say before sending a large invoice.”

“It does,” she said. “Unfortunately, it’s also true.”

She leaned forward slightly.

“Capacity is people. Execution architecture is how work moves from intent to outcome. Who owns it. How it is broken down. How decisions are made. How tools are used. How AI participates. How quality is checked. How accountability is measured. How fast you can expand or contract effort without rebuilding the company every time.”

John looked at his laptop again.

The blank section under Execution Capacity seemed less like a missing paragraph now and more like a confession.

For years, the company had treated work as something teams absorbed.

A customer needed a custom integration? Give it to engineering.

A reporting issue? Give it to data.

A broken process? Give it to operations.

A compliance demand? Give it to legal and product.

A strategic initiative? Create a cross-functional working group, which was corporate language for giving already busy people a shared calendar invite and no real capacity.

The company did not lack smart people.

It lacked a way to turn work into accountable outcomes without throwing each new demand into the same overloaded machinery.

That machinery had been heroic once.

Now it was congested.

Dinner trays were cleared. More passengers had surrendered to sleep. The cabin had entered that strange middle stage of long-haul travel where time felt suspended and everyone was temporarily removed from consequence.

John wished the company could feel this still for just one day.

No Slack. No escalations. No dashboards pretending to explain what people already knew. No performance theater. No meetings about meetings. Just enough quiet to ask:

How should work actually flow now?

Not in 2010.

Not in the old offshore era.

Not in the fantasy version of AI where tools magically absorb responsibility.

Now.

In a world where a company might have employees in London, customers in New York, engineers in Warsaw, contractors in India, compliance in Germany, a board in Boston, software everywhere, AI agents appearing inside workflows, and nobody fully sure who owns the final outcome.

John opened the document again.

This time, he did not write a slide title.

He wrote a question.

What work should no longer be trapped inside fixed teams?

Then another.

Where do we need outcomes, not roles?

Then another.

Which capabilities must become elastic?

Then another.

What can AI accelerate only if humans redesign the work around it?

He paused.

The questions made him uncomfortable.

Good questions often do. Bad questions flatter the existing model. Good questions threaten it.

Elena glanced at the screen.

“Better,” she said.

“You always this direct with strangers?”

“Only on night flights.”

He smiled.

For the next twenty minutes, they talked in low voices while the Atlantic passed beneath them.

Not about software features.

Not about productivity slogans.

About work.

About the strange way companies had become full of tools and still short of outcomes.

About the fact that many leaders were privately tired of the old choices: hire, outsource, delay, or overwork.

About AI, and why its biggest promise would be wasted if companies simply placed it on top of broken workflows.

About offshore delivery centers, and how they had solved one generation of problems but were not designed for a world of fluid work, intelligent agents, outcome-based accountability, and permanent cost pressure.

About the difference between owning people and owning capability.

That phrase stayed with John.

Owning people versus owning capability.

It sounded harsh at first. Then obvious.

A company did not need to own every person who contributed to its outcomes. It needed to own the system by which outcomes were defined, governed, delivered, verified, and improved.

That was different.

Much different.

By the time the first hint of morning touched the edge of the window, John had not slept.

But the panic had changed shape.

It was no longer a fog.

It was a problem.

Problems could be worked with.

The cabin lights rose slowly. Flight attendants moved through the aisles with breakfast trays. The man two rows ahead woke up, looked briefly confused by his own existence, and then immediately checked his phone.

The screen in front of John now said:

Time to destination: 48 minutes.

New York was close.

The meeting was closer.

John looked at his revised document. It was still incomplete. Still rough. Still not the grand answer the board probably wanted.

But the opening slide had changed.

It no longer said:

Q3 Operating Plan — Revised

It said:

From Headcount to Execution Capacity

Under it, he had written:

We cannot keep solving every execution problem by adding fixed structure. The next phase of the company must be built around elastic, governed, AI-enabled delivery capacity — where work is organized around outcomes, not just teams.

He read it twice.

It was not perfect.

But it was honest.

And in leadership, honest is often the first version of useful.

As the aircraft began its descent, Elena closed her folder and placed it in her bag.

“Good luck with your board,” she said.

John nodded.

“Thanks. I think I’ll need it.”

“You’ll need more than luck,” she said.

He laughed.

“Direct until landing.”

“Always.”

She stood to take her coat from the closet, then turned back.

“One more thing. When you get into that room, don’t start with AI. Don’t start with cost. Don’t start with hiring. Start with the work.”

“The work?”

“Yes,” she said. “Ask them what work still needs to be done after every plan, every cut, every tool, and every reorg. That is where the truth is.”

A few minutes later, the skyline appeared in the distance.

New York in the morning always looked like it had already been awake for hours.

John watched the city come closer and thought about the people in his company who were waking up across different time zones. London. Warsaw. New York. Austin. Bangalore. Berlin.

All of them carrying pieces of work.

All of them inside a structure that had made sense once.

All of them waiting, in one way or another, for the company to decide what kind of organization it wanted to become next.

The wheels touched down at JFK.

Phones came alive. Seat belts clicked. Passengers became executives again.

John opened his laptop one last time before closing it.

At the bottom of the first slide, beneath the title, he added one final line.

It was not polished.

It was not comfortable.

But it was the question he knew would define the meeting.

Maybe the company.

Maybe many companies.

We cut cost. We froze hiring. We bought tools. We talked about AI.
Now who delivers the work?

That was where the real conversation would begin.