The shift from people-centred to data-centred processes has happened quietly. But it changes everything: how processes must be thought through, documented and managed.
As long as a process was primarily carried out by people, it could live in their heads. Experience, intuition and informal agreements filled the gaps. That worked.
As soon as data, automation and AI enter the picture, that stops working. Systems cannot fill gaps with intuition. They need clear definitions: what comes in? What should come out? What happens in between?
That is the idea behind the IPO Framework.
A Model With History
The IPO model is not a new invention. In computer science it has been the foundational model for every data-processing system since the 1950s: a system receives an input, processes it, and produces an output. Simple, complete, universal.
What has changed are business processes.
Operational workflows that once ran purely manually, order processing, quoting, procurement, service planning, onboarding, are increasingly data-driven. ERP systems, automation tools, AI assistants and sensor technology are advancing into areas that a decade ago were managed exclusively by people and paper.
Business processes are therefore becoming structurally closer to computer science. They receive data, process it, and produce results. The logic is the same, only the context differs.
MAI DATA applies the IPO model consistently to operational business processes in industrial mid-sized companies, with one decisive difference from the classic IT perspective: the focus is not on technology, but on business value.
The Problem It Solves
Ask three employees in a company how a particular process works and you get three different answers. Not because anyone is doing a bad job. But because the process was never fully defined.
What gets documented are the steps in between. What is almost always missing: a clear definition of what must go into the process, and what should concretely come out at the end.
That has consequences. Everyone fills the gaps with assumptions. Handovers between departments fail because nobody has defined what needs to be handed over. Follow-up questions cost time. And when automation or optimisation is built on this basis, the problem does not get smaller, it gets faster and more systematic.
Automation does not solve a structural problem. It accelerates it.
The Three Sectors
The IPO Framework divides every process into three clearly defined sectors. It is not an IT concept, it is a tool for anyone who wants to understand, document and improve processes.
Execute from left to right.
Input — What Must Come In?
The input sector defines everything a process needs before it can be started. What data and information must be present, complete, of sufficient quality, at the right time?
In industrial mid-sized companies, data sources are often heterogeneous: ERP, Excel, email, paper and machine sensors coexist without a clear hierarchy. Every channel delivers information, but nobody has defined which of them are mandatory before the next step begins.
The central question: What information must be fully available before this process may start? As long as this question remains unanswered, every process runs on assumptions.
Processing — What Happens In Between?
The processing sector describes how the input is transformed into the desired output. This is where real value is created, and also where the greatest optimisation and automation potential lies.
Output — What Should Concretely Come Out?
The output sector describes the concrete result of the process. Not abstract, but precise: what exists at the end? For whom? In what format? By when?
A well-defined output enables the right person to make the right decision at the right time, without further preparation. That is the benchmark.
The Application Logic: Output-first
The IPO Framework is not applied from left to right. The logic is reversed, and that is the decisive difference from classical process documentation.
Define the output
What should concretely exist at the end, and what business value does it have? For whom? In what format? By when? A good output is SMART.
Clarify the input
What data and information does the process need to produce this output? What is mandatory, what is optional?
Structure the processing
Which steps transform the input into the desired output, and which of them can be automated?
Thinking from the output reveals things that would otherwise be missed. Which data must be collected from the start. Which handover should have happened earlier. Which step is not necessary at all because it contributes nothing to the defined output.
Practical Example: Quoting in Mechanical Engineering
A mechanical engineering company receives a customer enquiry. The calculation begins, somehow. Three days later the quote goes out. Too late. The customer has already decided to go elsewhere.
The IPO Framework makes it visible why.
The result: quotes that go out faster, contain fewer errors, and customers who do not switch to a competitor because the response took too long.
What the Framework Achieves, and What It Does Not
The IPO Framework creates clarity. But clarity alone does not change anything. The next step is optimisation, then automation, then intelligent knowledge management that ensures the knowledge developed stays in the organisation.
What the framework achieves: it ensures that no process is built on assumptions anymore. That input and output are defined so clearly that automation is not built on sand. And that three employees describing the same process give the same answer.
In the following articles we go through each sector in detail, what makes good input, where in the processing the most potential lies, and what output definitions look like in practice that actually work.