Digitalising processes: the right order of steps

The right order for digitalising processes is: simplify, then digitise, then automate, then AI. Underneath, there has to be stable IT infrastructure: network, Wi-Fi, backups and access rights. Each stage builds on the one before. Anyone who starts with AI, or digitises a cumbersome process as it is, rebuilds the existing detours in software and keeps paying for them every day.

Why does the order of steps matter in digitalisation?

Every stage of digitalisation amplifies what it finds. Software makes a clear process faster and more reliable. Software also makes an unclear process faster, but harder to change, because the detours are now locked into code, forms and interfaces.

The same applies one stage higher. Automation running on messy data produces errors without human oversight. An AI that draws on scattered and contradictory documents gives answers that sound convincing and are still wrong. Whether an investment pays off therefore depends on the order.

What needs to be in place before you digitalise processes?

Before a business digitalises its workflows, the IT infrastructure has to hold up. A tablet in the warehouse is useless if the Wi-Fi drops out on the shop floor. Digital time tracking on a construction site fails if logging in does not work reliably. And new software is a risk if its data is not backed up.

• Network and Wi-Fi work where people actually work, including the shop floor, warehouse and workshop.
• It has been decided which data lives on your own servers and which in the cloud.
• Backups run and a restore has actually been tested.
• Every person has their own user account with appropriate access rights.
• Devices and programs receive security updates regularly.
• There is up-to-date documentation of who looks after which system.

This foundation is unspectacular, but everything else rests on it. In practice it often turns out that a digitalisation project does not fail because of the software, but because of an access point mounted in the wrong place.

Stage 1: Why simplify first?

Simplifying means looking at a process at the workstation and dropping everything that no longer serves a purpose: duplicate approvals, lists nobody reads any more, printouts that only get filed. This stage needs no software and still often brings the most. How to map a process at the workstation and find the bottleneck is something we describe in our article on process optimisation for SMEs.

At the same time, this stage makes clear who needs which information in the process and where it originates. That knowledge is exactly what you need later to choose the right software or build the right interface.

Stage 2: What does digitising properly mean?

Digitising means a piece of information is captured digitally once, where it originates, and then moves on without being retyped. A scanned paper form is not yet digitalisation as long as someone has to copy the values out of it again. A process is properly digitised when goods receipt is booked in the warehouse, the service report is completed at the customer's site and working hours are entered on the construction site.

This stage also shows whether existing programs are enough. Inventory, accounting and time-tracking systems can often do more than they are used for, and only the connection between them is missing. An interface between existing systems is frequently more sensible than yet another isolated tool.

Stage 3: When is automation worth it?

Automation is worth it when a step happens frequently, always follows the same rules and the data it needs is already digital and reliable. The software can then take over that step, and employees deal with the exceptions.

• Once an order is completed, the invoice is prepared automatically.
• When an item drops below minimum stock, a purchase proposal is created.
• Maintenance is scheduled automatically when a machine reaches its service interval.
• New customer data from the online shop reaches the ERP and accounting without retyping.

Every automation needs monitoring and someone who maintains it. If an automatic transfer silently fails, nobody notices without a check until a customer asks.

Stage 4: When is AI the right next step?

AI is the right next step when tasks remain that cannot be captured in fixed rules: understanding and pre-sorting enquiries from emails, finding the right passage in manuals and inspection records, drafting quotation texts from existing documents. An AI assistant is only as good as the documents it can access. That is why AI comes at the end of the sequence, not at the beginning.

AI assistants do not necessarily have to run in the cloud. They can also be operated locally on your own hardware, so customer data and internal documents never leave the building. With regard to the GDPR and the EU AI Act, it should be clarified before deployment which data is processed and who reviews the results. This does not replace a case-by-case legal assessment.

How can you tell which stage a process is currently at?

The stage of a process can be assessed with a few observations at the workstation. What counts is what people actually do, not which software is licensed. A business can own a modern ERP and still be at the first stage for many processes, because people keep working on paper and in Excel alongside it.

• Before stage 1: nobody can describe the process completely, and everyone does it slightly differently.
• Stage 1 reached: the process is clear and unnecessary steps are gone, but information still travels on paper or by retyping.
• Stage 2 reached: information is captured digitally once and passed on without retyping.
• Stage 3 reached: rule-based routine steps run automatically, and people handle the exceptions.
• Stage 4 makes sense: tasks requiring judgement remain, and reliable digital documents are available for them.

This assessment is not a verdict on the business but a working basis. It shows which step brings the most benefit next and prevents budget from flowing into a stage whose prerequisites are still missing.

What does the sequence look like as a concrete roadmap?

1. 01Check the foundation: get network, Wi-Fi, backups, user accounts and updates in order.
2. 02Choose one process that happens often and clearly causes frustration.
3. 03Map the process at the workstation and drop every step that no longer serves a purpose.
4. 04Capture information digitally at the source and eliminate retyping between programs.
5. 05Connect existing programs via interfaces before buying new software.
6. 06Automate recurring, rule-based steps and monitor the automation.
7. 07For remaining tasks without fixed rules, check whether an AI assistant makes sense, run locally if needed.

What happens if you skip a stage?

• Without simplification, the detour gets digitised and is then harder to change than on paper.
• Without clean digitisation, automation works with incomplete or wrong data.
• Without automating the routine, AI ends up doing tasks a simple rule would handle better.
• Without stable infrastructure, digital processes fail exactly when they are needed.

How NDVDL takes over digitalisation in your business

NDVDL covers every stage, from IT infrastructure through software and interfaces to AI development. That does not mean we sell everything at once. We come to your business, let the people who do the work show us the processes, and then tell you in writing which stage a process is at and which next step is worth taking. Sometimes that is a new access point, sometimes an interface, sometimes an AI assistant.

Implementation happens step by step while your business keeps running. Afterwards NDVDL operates and maintains the infrastructure, software and automations, with one fixed contact person. You do not have to coordinate between a network technician, a software house and an AI vendor.