We are entering an era in which increasingly sophisticated work can be delegated to AI agents. These systems can plan and execute multistep workflows with limited supervision.
That is exciting, but it also creates a real leadership risk.
Organizations can now automate processes that once required significant time, money, and technical expertise. But just because we can automate something does not mean we should.
I am not approaching this as an automation skeptic. I am heavily using agentic platforms such as Hermes, Grok Bot, and Manus to execute sophisticated workflows. In some cases, these are long-horizon tasks that continue working over hours or days.
These tools are rapidly expanding what leaders and organizations can accomplish. That is exactly why we need to become more disciplined about what we ask them to do.
When execution becomes easier, judgment becomes even more important.
The Most Efficient Waste Is Still Waste
Peter Drucker captured the problem decades ago:
“There is surely nothing quite so useless as doing with great efficiency what should not be done at all.”
Automation can make a good process faster, more consistent, and more scalable. It can do the same thing to a bad process.
Automating unnecessary work does not eliminate waste. It institutionalizes it.
We can end up with faster reports that nobody needs, more frequent notifications that nobody reads, or sophisticated workflows built around requirements that should have been reconsidered years ago.
Our Instinct Is to Add
When leaders encounter a problem, our instinct is often to add something.
We add another tool, approval, meeting, report, step, or layer of technology. Adding feels like progress because something visible has changed.
Subtraction requires a different kind of discipline.
A 2021 study published in Nature found that people systematically defaulted to additive changes and often overlooked beneficial subtractive solutions. Participants considered subtraction more frequently when they were specifically prompted to think about it.
That matters for leaders.
Subtraction rarely happens by accident. We must deliberately ask what can be removed, simplified, or stopped.
Some of the greatest organizational gains come not from adding another capability, but from removing work that no longer deserves to exist.
Return to First Principles
Before asking how to automate a workflow, break it down to First Principles.
What outcome are we actually trying to create? What is truly necessary to produce that outcome? Which parts of the current process exist only because of history, habit, or an outdated constraint?
The Five Whys is a simple root-cause analysis technique. State the problem, ask why it exists, and then continue asking why after each answer. Five is not a magic number. The point is to move beyond the visible symptom until you uncover the underlying need, assumption, or constraint.
Applied to automation, the first answer may explain how the workflow operates. The next few answers may reveal why it was created, whether that reason still exists, and whether the workflow remains the best way to produce the desired outcome.
The goal is not to defend the process we inherited. It is to understand the problem well enough to rebuild the right solution.
Do not begin with:
How can we automate this?
Begin with:
Why does this exist?
Before You Automate, Ask Three Questions
Before investing time, technology, or attention in automation, ask:
If we were starting today, would we design the workflow this way?
Separate genuine requirements from inherited habits.Does this work still create meaningful value?
If it disappeared tomorrow, who would notice and what would suffer?What can be eliminated, simplified, combined, or redesigned first?
Force subtraction onto the table before discussing technology.
Automation Comes Last
Once the workflow has survived that audit, automation can become an extraordinary force multiplier.
Before implementation, it must also satisfy the organization’s security, privacy, AI governance, and IT governance requirements. Leaders should know what data the workflow can access, which actions it can take, where human approval is required, and who remains accountable for the outcome.
Eliminate what no longer matters. Simplify or redesign what still creates value. Delegate what belongs with someone else. Then automate the stable, secure, and repeatable work that remains.
The goal is not to automate more work. It is to ensure that human and technological effort is spent on work worth doing.
This week, choose one recurring workflow that your organization is considering automating. Before making it faster, run it through the three questions.
You may discover that the most valuable improvement is not something you add.
It is something you remove.
Here’s to eliminating before automating,
Steve Kahle | ELITE OPERATORS
In Pursuit of Elite
