Chasing Immediate Progress Can Make Work Slower: The AI Era Demands Time to Sharpen the Axe
When an IT development project begins, the priority is usually to clear the tasks immediately in front of us.
Close tickets. Build screens. Implement APIs. Push the progress figure up, even if only slightly.
There is nothing inherently wrong with that. But working with AI lately has reminded me of something.
Rather than pressing ahead with the immediate work, wouldn’t we ultimately accomplish far more by first creating an environment that makes the work faster?
Make the AI Smarter Before Writing Code
Recently, I have been spending a considerable amount of time improving my Claude Code development environment.
I organize project-specific rules and give the AI the right context. I establish review procedures and automate recurring tasks.
While I am doing this, no actual features are being developed.
If I spent that time writing code instead, my immediate progress would certainly be more visible.
Once the system is in place, however, it improves every task that follows.
For example, if I consolidate rules that I have been explaining to the AI every time, I will no longer need to repeat those explanations. If I organize the review criteria in advance, I will not need to rethink them from scratch each time.
The saving on any single task may be small, but repeat it dozens or hundreds of times and the difference becomes substantial.
People have long said that you should sharpen the axe before cutting down a tree. This is exactly what they mean.
Why Does This Work Always Get Postponed?
And yet, the busier a development team becomes, the more likely it is to postpone this kind of groundwork.
The reason is simple.
People get in trouble if they do not finish the work immediately in front of them.
When a feature must be implemented by the end of the week, reporting, “I spent today improving our AI development environment,” is unlikely to be met with much understanding.
On a project that is already behind schedule, someone may well say, “If you have time for that, write some code.”
The result is that the team keeps repeating inefficient work indefinitely.
They are too busy to improve efficiency. And because they never improve efficiency, they remain busy forever.
It is an exceptionally straightforward vicious cycle.
Project Managers Should Make Time for It
It is difficult to solve this situation through the efforts of individual engineers alone.
That is precisely why project managers and team leaders need to include time for improving the development environment in the plan from the outset.
Setting up the development environment, configuring the AI, establishing shared rules, and building automation.
These should be treated as formal development tasks, not as work to be done “if there is time.”
Teams also need regular opportunities to reconsider how they work.
Even if it is only for a short time each week, they should consider whether their current tasks could be made easier and whether any part of the work could be delegated to AI.
These improvements affect not only individual productivity but also the development speed of the entire team.
Of course, improving the environment must not become an end in itself. There is no value in overengineering automation tools that nobody uses.
What matters is identifying work that occurs repeatedly and investing in it upfront.
Work That Does Not Appear in the Progress Figures
Traditional project management inevitably tends to reward completed features and closed tickets.
In the AI era, however, the value of someone who creates a system that allows the same kind of work to be completed faster again and again is greater than ever.
Completing one feature, or making the development of every future feature faster.
It is obvious which one will produce greater results over the long term.
The problem is that the benefit does not immediately appear in the numbers.
That is why project managers must look beyond immediate progress and make decisions with future productivity in mind.
Some teams begrudge the time needed to sharpen the axe and keep swinging a blunt one instead.
Workplaces like that have always existed.
If a team is still swinging an axe and shouting, “Cut faster!” even after AI has given us a power chainsaw, then the problem is no longer the tool.