The request sounded straightforward.

"We want to automate this with AI."

Good.

What does the process look like today?

A pause.

Then the explanation started.

First, one person receives the request.

Sometimes by email.

Sometimes by phone.

Sometimes through a form.

Then information gets copied somewhere else.

Unless something is missing.

Then somebody asks for it.

Another person checks the request.

Sometimes.

Certain cases require approval.

Although nobody was completely sure which ones.

Then the information moves to another team.

Except when it doesn't.

And several people had slightly different explanations of what happened next.

The AI wasn't the first problem.

The process was.

Automation sounds like improvement

And sometimes it is.

AI can now perform work that would have required enormous amounts of human effort only a few years ago.

Read documents.

Summarize information.

Classify requests.

Generate responses.

Analyze data.

Support customers.

Write code.

Trigger workflows.

The opportunity is enormous.

But there's an uncomfortable truth:

Automation doesn't automatically improve a process.

Sometimes it simply executes the existing process faster.

So we asked them to show us the process

Not describe it.

Show it.

Where does work enter?

What information is required?

Who touches it?

What decisions are made?

Where does it wait?

What happens when something goes wrong?

Which exceptions exist?

What does the customer experience?

Suddenly the AI conversation became much less important.

Because before automating anything, there were basic operational questions that didn't have consistent answers.

Different people were solving the same process differently

That's dangerous.

Not necessarily because people were doing bad work.

Quite the opposite.

Experienced employees had learned how to compensate for weaknesses in the system.

They knew who to ask.

Which shortcut to use.

Which customer needed special treatment.

Which information could be ignored.

Which mistake needed checking.

Their experience was holding the process together.

And now we were considering teaching a machine to execute it.

This is where AI can scale the wrong thing

Imagine a process contains unnecessary steps.

AI makes them faster.

There are poor decisions.

AI helps make more of them.

Information enters incorrectly.

AI distributes it more efficiently.

The customer journey is confusing.

AI allows more customers to experience it.

That's not transformation.

That's acceleration.

And acceleration is only useful when you're moving in the right direction.

So the question changed

Instead of:

"Where can we use AI?"

We asked:

"If we were designing this process today, would we build it this way?"

Now we're having a transformation conversation.

Which steps actually create value?

Which decisions require humans?

Which decisions can follow rules?

What information is genuinely necessary?

Which exceptions should disappear?

Where does judgment matter?

Where does automation create leverage?

Only after answering those questions does AI become interesting.

Process first. Technology second.

This doesn't mean spending six months documenting everything before touching AI.

Quite the opposite.

The objective is to simplify enough that technology amplifies something worth scaling.

Sometimes the answer will be AI.

Sometimes ordinary automation.

Sometimes a clearer responsibility.

Sometimes removing three steps.

Sometimes simply stopping something entirely.

The technology should follow the problem.

Not the other way around.

Then comes execution

Let's say the redesigned process is clear.

Three steps disappear.

Two responsibilities change.

One activity becomes automated.

A new AI workflow is introduced.

Success metrics are defined.

Great.

But transformation still hasn't happened.

Someone needs to own each change.

Someone needs to test it.

Someone needs to know when it fails.

Someone needs to measure whether the new process is actually better.

Someone needs to follow up.

Otherwise...

You've simply created a smarter-looking version of the old system.

AI doesn't remove leadership

It increases the importance of it.

Because the easier execution becomes, the more important the decisions behind execution become.

AI can help do the work.

It can't decide what your organization should become.

That's still the job.

One question worth asking

Before your next AI initiative, ask:

"If AI didn't exist, would this process still make sense?"

If the answer is no...

Don't automate it yet.

Fix what you're asking the technology to scale.

INSIDE THE BUSINESS

Real problems. Hidden causes. Strategic questions.

At Bozhi-Dar Strategic Consulting, we help businesses understand the real problem before choosing the solution, redesign how work should flow, and create the ownership and accountability needed to turn transformation into reality.

Technology creates leverage.

But leverage multiplies whatever is already underneath it.

If you're considering AI but aren't completely sure the underlying process deserves to be automated, let's have a conversation.