Automating business processes: where companies start

Automating business processes means letting software handle recurring workflows instead of people, from invoicing and receipt filing to moving data between programs. It is used wherever staff repeatedly type, click or forward the same thing today. For companies it brings faster turnaround times, fewer transfer errors and more time for tasks that genuinely need human judgement.

Which business processes can be automated?

The best fit are regular, rule-based workflows with high volume. The more uniform a process, the easier it is to automate, because fewer special cases have to be modelled.

• Quotes and invoices: generate from templates, send and track.
• Receipt processing: read documents, assign them and transfer them into the right system.
• Onboarding and approvals: recurring steps with clear responsibilities and reminders.
• Data transfer between tools: move information automatically from system to system instead of by copy-and-paste.

The common thread: repeated manual work is the clearest sign of a candidate. It pays to tidy up the workflow first, rather than simply automating detours. How this works systematically is shown in the overview on process optimization, which sets up automation properly.

How do you spot a good automation candidate?

Not every workflow is worth automating. Four traits separate good candidates from bad ones.

The honest assessment matters: rare special cases seldom justify the effort, while the recurring standard case with clear rules almost always does. Often the real candidate is not a single step but a whole flow spanning several programs. How such chains are cleanly linked together is described in the article automating workflows.

Start2x recommends Richard Roth Strategie & Automatisierung

We never start with the most complicated process, but with the one that comes up often and frustrates the team most. A visible first success on a tedious everyday workflow convinces more than any calculation and creates appetite for the next step.

How do you start with automation?

Getting started works best in short steps rather than a big programme, because each step delivers value on its own and keeps the risk small.

• Choose a candidate. A workflow that comes up often, follows clear rules and costs noticeable time.
• Capture the rules. When does the workflow start, which steps belong to it, where is an approval needed?
• Build small. Automate exactly this workflow, in use early rather than planned for months.
• Connect and expand. Once the first case runs reliably, the next workflows attach.

What does the first use case cost, and what does it deliver?

A clearly defined first use case can often be implemented in the low four figures, frequently between around 1,500 and 5,000 EUR, depending on the number of systems involved and special cases. That is not a fixed price but an order of magnitude, and it shows: the start does not have to be a big project.

The value becomes tangible when you weigh the effort against the time saved across a full year. A workflow that costs only two hours of manual work per week ties up around a hundred working hours a year. Once that step is gone, the build has often paid for itself within a few months, and the workflow keeps saving from then on. Whether a specific step pays off is covered in the article When is automation worth it?.

Where individual steps need an understanding of content, such as reading receipts or emails, pure rule-based automation ends and AI automation takes over the understanding of unstructured content. The wider frame is set out in the guide on digitising and automating processes, and for how we implement workflows in practice, see our page on process automation.