Processing: where value
is created and potential
is lost

Processing is where the real value is created. And it is where most of the potential is lost, because it is the part everyone feels but no one sees.

The part everyone feels but no one sees

Everyone in the company knows the feeling. A task that drags on. An approval you wait for. A piece of information you request for the third time. Something that has always been done this way, without anyone being able to say why.

That is processing. The part between input and output where the actual work happens. This is where value is created. And this is where most of the potential is lost.

The tricky thing: in the input you can name a problem, missing data, wrong values. In the output too, the result is off. But processing is diffuse. It lives in the heads of the people who carry it out every day. In informal arrangements. In habits no one questions anymore. That is exactly why it is the sector with the greatest room for improvement and the least understanding.

The order that decides everything

When a process is stuck, the first reflex is often: automate it. A piece of software, a tool, a bit of AI, and the problem is solved.

It is the most expensive reflex there is.

Because the right order in processing is: understand, optimise, automate. In exactly that sequence. Skip a step and you pay double later.

Step 1: Understand

Before improving anything, you have to make the process visible. Not the way it appears in the org chart or how the manager describes it, but the way it actually runs.

A simple distinction helps: which steps create real value for the result, and which exist only because they grew historically? The question sounds trivial but is uncomfortable. Because almost every process contains steps that no longer serve anyone. An approval no one reads anymore. A form filled in twice. A report no one needs.

On top of that come the invisible losses. The waiting times where a task simply sits. The handovers where information gets lost. The duplicate work because two departments capture the same thing. In many processes, the actual handling time is only a fraction of the total lead time. The rest is spent waiting.

Whoever makes this visible almost always finds more than expected. And often the biggest improvement is not making something faster, but leaving something out.

Step 2: Optimise

Only once the process is understood is optimisation worthwhile. And it follows a clear principle: first remove, then improve, then and only then automate.

Removing means leaving out everything that creates no value. It is the most effective and cheapest lever there is, because the best process step is the one that no longer exists.

Improving means smoothing the remaining steps. Shortening waiting times, defining handovers cleanly, resolving bottlenecks. A process is only ever as fast as its slowest point, and finding that one point brings more than speeding up ten others.

This is where many projects make their mistake. They skip optimising and go straight to the technology.

A bad process that gets automated is a fast bad process. The errors don't disappear, they just happen more often and at a larger scale.

Step 3: Automate

Now, and only now, technology comes into play. An understood, optimised process can be automated deliberately, because it is clear what should be automated and why.

But here too it takes honesty. Not everything should be automated. There are steps that are rule-based and repeatable, clear candidates for automation. And there are steps that require judgement, experience or human assessment.

Deliberately keeping these steps with people is not a weakness of automation, it is good design. The human in the process, in the right place, is not a transitional state until full automation. It is often the best solution. The art lies in knowing which step needs what.

And where does AI fit in?

In processing, AI can do things rule-based automation cannot. Classify unstructured data, recognise patterns, process language. That opens up real possibilities.

But the principle stays the same. Putting AI on a process that is not understood and not optimised only scales the chaos instead of solving it. Structure first, then intelligence.

That is exactly why processing is not the sector where you start with technology. It is the sector where you start with understanding and bring in technology as the last, deliberate step.

What remains

Processing is where it is decided whether a process creates value or produces waste. And the most important insight is almost always the same: the greatest potential lies not in more technology, but in less complexity.

Whoever understands their process before optimising it, and optimises it before automating it, builds on solid ground. The next article is about the final sector, the output. And about perhaps the most important question of all: how do you actually tell that a process delivers the right thing?