The New Bottleneck: When Thinking Becomes the Real Constraint

For a long time, organisations were constrained by execution.

Getting things done was:

• slow
• expensive
• unpredictable

That’s why we built systems around it:

• roles to divide work (as explored in The Death of Roles)
• planning to reduce uncertainty (see Why Planning Breaks When Execution Stops Being the Problem)
• workflows to coordinate effort (as described in Workflows Without Humans)

All of it made sense.

Because execution was the bottleneck.

What Happens When Execution Speeds Up

Now execution is changing.

Tasks that used to require:

• deep expertise
• significant time
• multiple iterations

Can now be:

• drafted instantly
• refined quickly
• repeated consistently

This doesn’t remove the need for people.

But it changes where the friction is.

When execution becomes easy, it stops being the limiting factor.

The Bottleneck Moves

And when one bottleneck disappears, another one becomes visible.

In this case:

The bottleneck shifts from doing → to thinking.

Not thinking in the abstract sense.

But very concrete capabilities:

• framing the right problem
• asking the right questions
• making meaningful trade-offs
• deciding what actually matters

These are not easily automated.

And they don’t scale the same way execution does.

Why Thinking Doesn’t Scale Like Execution

Execution can be:

• standardised
• automated
• optimised

Thinking cannot.

It is:

• contextual
• ambiguous
• dependent on judgment

You can support it.

You can accelerate parts of it.

But you can’t fully delegate it.

And that creates a new kind of constraint:

Not a lack of capacity—but a lack of clarity.

More Output, Less Direction

This is already visible in many teams.

With AI, it becomes easier to:

• generate ideas
• produce content
• explore options

But that creates a new problem:

More output requires better filtering.

If you can create 10 options in minutes, then:

• which one do you choose?
• what criteria do you use?
• what are you optimizing for?

Without clear thinking, more output doesn’t create more value.

It creates more noise.

The Illusion of Progress

This is where things get tricky.

Because high output looks like progress.

• more tickets completed
• more features delivered
• more content created

But if the direction is unclear:

You can move faster in the wrong direction.

And AI makes that easier than ever.

Where This Shows Up in Organisations

You start to see new patterns:

• teams delivering quickly, but without clear impact
• decisions being delayed, even though execution is fast
• endless options, but no alignment

It’s not a delivery problem anymore.

It’s a thinking problem.

From Doing More to Deciding Better

This requires a shift.

From valuing:

• productivity
• efficiency
• output

To valuing:

• clarity
• judgment
• decision quality

Not because the first doesn’t matter.

But because the second becomes the constraint.

What Becomes Valuable

If thinking is the bottleneck, then different capabilities gain importance.

Not:

• how fast you execute
• how much you produce

But:

• can you simplify complexity?
• can you challenge assumptions?
• can you define what success looks like?
• can you say no to good ideas?

These are not new skills.

But they become more visible—and more critical.

Where This Gets Personal

I notice this shift in a very practical way.

Tasks that used to take significant effort now feel almost trivial with AI support. Drafting, structuring, even exploring different angles—it’s all faster.

But instead of freeing everything up, it creates a different kind of tension.

Because the question is no longer how to do something.

It’s whether it’s worth doing at all.

And I find that this question takes more time than the execution ever did.

Not because it’s complicated in theory.

But because it forces you to be clear.

Clear about:

• what you’re trying to achieve
• what you actually believe
• what you’re willing to ignore

And there’s no shortcut for that.

The Organisational Gap

Most organisations are not designed for this shift.

They are still optimised for:

• managing execution
• tracking output
• increasing efficiency

But they struggle with:

• improving decision quality
• enabling better thinking
• creating real clarity

So the system keeps pushing for more output…

While the real constraint sits elsewhere.

Closing Thought

For years, the challenge was getting things done.

Now the challenge is deciding what should be done.

AI doesn’t remove the need for people.

It removes the excuse for not thinking clearly.