Key Takeaways
The Gap Between the Designed Process and the Real Process
Companies have never had more technology. ERP systems. Dashboards. Workflow tools. Automation platforms. AI assistants. And yet, walk into many operational teams and you will find something remarkably familiar: Excel sheets. Emails. Manual follow-ups. Information stored in people's heads.
For Max Brunner, Head of Vendor Management & Brand Relations at Kaufland e-commerce and founder of Startforge AI, this contradiction points to a specific and underappreciated problem. Perhaps the issue is not primarily the technology. Perhaps we simply don't understand the last 20% of the operation well enough.
Max immediately recognised a distinction that sits at the core of this problem: the gap between what product managers and developers assume a user will do — and what the user actually does. People aren't machines. They take shortcuts. They forget things. They use whatever tool is easiest at that particular moment.
"The process says information should be entered into the system. Reality says someone sends an email."
The process says a status should be updated. Reality says somebody calls the supplier. The process says responsibilities are clearly defined. Reality says three people are copied into an email and everybody assumes someone else will handle it. Eventually the system gets updated — but the operation happened somewhere else.
That is, in many ways, why Still On Email has its name.
The Last 20% Can Break the Entire Process
Transformation projects are often designed by people who aren't deeply embedded in the operation itself. They understand the major process steps. They understand perhaps 80% of what needs to happen. But it is the remaining 20% — the exceptions, informal knowledge, workarounds and small operational decisions — that determines whether the process actually works.
Take Purchase Order Management. On paper, it looks straightforward. Order placed. Supplier confirms. Production happens. Goods are booked. Container planned. Shipment departs. Cargo arrives. Everything fits neatly into a diagram.
But operations doesn't live in those boxes. A supplier doesn't respond. Production is delayed. A document is missing. A booking hasn't been confirmed. A container isn't fully utilised. Somebody changes a quantity. Someone else has information that hasn't reached the system yet.
Suddenly the operation depends not on the beautifully designed process — but on people knowing what to do next. That last 20% is where a huge amount of operational work actually lives.
Transformation Has to Reach Hans and Peter
Max is sceptical of the idea that a company creates one AI pilot team, runs a few impressive projects and then calls itself transformed. Instead, he argues that employees across the entire organisation need to understand what the technology can actually do.
The people closest to the operation are often the people best positioned to identify where technology creates real value. Management can identify the big rocks. But the operational team knows where twenty minutes disappear every morning. They know which Excel file constantly needs updating. They know which supplier needs chasing. They know which information exists only in somebody's head.
"Transformation hasn't really arrived if the project looks impressive while the relevant information is still sitting with Hans and Peter somewhere in the operation."
That may be one of the simplest tests for digital transformation: Did the technology change the presentation — or did it change the work?
From Dashboards to Decisions
More information does not automatically create better operations. Max sees this clearly from his work in Vendor Management. A good Vendor Manager isn't simply someone who can read a KPI dashboard. The real skill is translating that dashboard into an understanding of what should be done — and then convincing partners to act accordingly.
A dashboard can tell you that a Purchase Order is late. Operational control should help you understand why it is becoming late, what needed to happen before that point, who is responsible and what action might still protect the outcome.
Visibility tells us what happened. Good operations require us to decide what happens next.
AI Might Finally Change the Interface
For decades, enterprise software has largely asked humans to adapt to software. Fields must be completed in a particular way. Processes must follow predefined workflows. Users must learn the system. When reality doesn't fit neatly into the software, people create another Excel sheet or return to email.
Max sees large language models potentially changing that relationship. Instead of forcing every user through the same rigid dashboard or interface, AI creates the possibility of an interface that adapts much more closely to how the individual actually works.
"Perhaps the future isn't about eliminating email, Excel or every informal interaction. Perhaps it is about building systems capable of understanding the work happening around them."
An email arrives from a supplier. A production delay is mentioned. Instead of requiring somebody to manually transfer that information through three systems, technology understands what changed, which order is affected, what that means for the operational goal — and helps determine what needs to happen next.
The human doesn't disappear. The administrative friction does.
AI Should Not Remove Responsibility
Max doesn't describe AI as a replacement for thinking. Quite the opposite. His experience is that AI has dramatically shortened the time required to create and execute — but that changes where human effort goes, not whether it is needed.
The relative amount of time spent checking, questioning and critically evaluating becomes greater, even while the overall project moves much faster. That is an important correction to much of today's automation narrative.
Less copying data. Less checking whether somebody replied. Less manually updating status fields. Less chasing information across systems. And more time understanding exceptions, talking to partners, finding solutions and making real decisions.
From Technology Projects to Operational Transformation
Perhaps one of the biggest mistakes in digital transformation is starting with the technology. Max takes almost the opposite approach in his own projects. He starts with a problem. Then he defines what success would look like. Only afterwards does he work backwards towards the tools, processes and software required to achieve it.
That sequence matters. Operations doesn't need another AI project simply because AI is available. It doesn't need another dashboard simply because dashboards are possible. And it certainly doesn't need another system that employees have to maintain alongside the work they are already doing.
It needs technology that understands what people are trying to accomplish — and helps them get there.
Operations Is Where Transformation Becomes Real
After more than 15 years around logistics and supply chain operations, the pattern is hard to ignore: too many processes held together by people manually entering data, checking updates, maintaining Excel sheets and chasing information through email.
The next generation of operational technology shouldn't simply digitise the process drawn on a whiteboard. It has to understand the process that actually happens.
"Including the awkward last 20%. Including the email. Including the Excel sheet. Including the exception nobody planned for. Including Hans and Peter. Because that is where operations lives."
And perhaps that is also where digital transformation finally has to go.
Still On Email — Operations Unfiltered. Conversations about logistics, supply chain, operations, technology and the reality behind digital transformation.