6 November 2026 · 6 min read · Valendis Team

Automate Business Processes: Where to Start in a Small or Mid-Sized Company

Project manager and business owner at a whiteboard deciding which business process to automate first
The best first candidate is rarely your most important process. It is the most frequent one with clear rules.

Short answer: when you set out to automate business processes, do not start with your most important process. Start with the one that repeats most often, follows clear rules and eats the most hours today. Work out what three to five candidates cost you per year, run a small pilot, and measure how long the work takes before you begin. Within a few weeks you will know whether the investment pays off.

Many companies stall at exactly this point. According to the KfW-Digitalisierungsbericht Mittelstand 2025 (German, PDF), only 30 % of German SMEs completed a digitalisation project between 2022 and 2024, with a lack of IT know-how and financing among the hurdles. That is why the first project should be cheap and show results quickly. This guide helps you make that call in an afternoon.

Five signs a process is ready for automation

  • People type the same data into two or more programs.
  • Lists and spreadsheets are updated and sent out by hand every week.
  • Customer enquiries wait because nobody is clearly responsible.
  • Reports appear at month end, when it is too late to correct course.
  • Mistakes are only noticed by the customer or by accounting.

If one of these fits a task that happens every week, you have a candidate.

Employee copying figures from paper invoices into a spreadsheet by hand, a typical process to automate

Step 1: Collect three to five candidates

Ask the people who do the work every day, not only management: "Which task would you most like never to do again?" Their answers are usually sharper than any process map. Typical candidates in small and mid-sized companies:

  • Capturing enquiries from the website, email and phone and passing them to the right person
  • Building quotes from templates and price lists
  • Checking incoming invoices, matching them and sending them for approval
  • Pulling weekly and monthly reports together from several sources
  • Confirming appointments and sending reminders
Team collecting recurring tasks on sticky notes in a workshop to find processes worth automating

Step 2: Put a yearly cost on each one

A simple formula is enough: hours per week × fully loaded hourly cost × working weeks per year.

Notepad with a scoring grid, calculator and timer used to calculate the yearly cost of a business process

A sample calculation: a project manager spends six hours a week on quotes. At €45 an hour and 46 working weeks, that is €12,420 a year. If automation saves half of that time, you have more than €6,000 a year to weigh against a one-off cost. A number like this makes the decision concrete, for partners and for the bank.

Add the cost of mistakes too, such as forgotten follow-ups or prices copied wrongly. They are harder to put a figure on, but often the bigger item.

Step 3: Score the candidates and pick your pilot

Give each candidate 1 to 3 points in five categories:

Criterion1 point3 points
How oftenmonthlydaily
Time spentunder 2 h/weekover 5 h/week
Clear rulesmany exceptionssame steps every time
Data already digitalpaper, handwritten notesemail, forms, spreadsheets
Cost of mistakeslowlost money or customers
Sticky notes placed in a priority matrix to choose the first process to automate

The highest score becomes your pilot. A common mistake is to start with the most strategic process: it usually has the most exceptions, the most people involved and the highest risk. A quick, visible win does more. It convinces the team and funds the next step.

What not to automate first: processes that are changing right now, decisions that need a lot of judgement, and anything nobody can fully describe. These need clarity first, technology second.

Step 4: Start small and measure first

Before you start, write down how things work today: hours per week, time from request to answer, and errors per month. Without this baseline you cannot show later what automation achieved.

A well-chosen pilot goes live in two to six weeks. Run the old and new process side by side for a while before you switch. That way you find the special cases before your customers do.

An example from our own work: on Alpin Sun's 89.4 MW solar park build, progress for 4,228 module tables was copied by hand into one large Excel workbook. That meant late numbers, typos and hours of office work every day. The daily report happened every day and followed fixed rules, an ideal first candidate. Today it arrives automatically at 7:00 every morning, and the team still works in the workbook it knows. Read the Alpin Sun case study and our article on automating construction progress reporting.

Before, keeping the report up to date by hand took about four hours a day. Over roughly 20 working days, that is around 80 hours a month now free for other work.

Construction worker recording solar park progress on a tablet, with satellite internet on the site container behind

Step 5: Bring your team with you

Automation rarely fails because of technology; it fails because people do not accept it. Say early and openly which tasks will go away and what the freed-up time is for. Make the person who does the task today your tester. They know the special cases best.

Manager explaining the new automated workflow to her team on a laptop

If the process touches customer or employee data, bring data protection under the GDPR in from day one, and check whether employee representatives need to be consulted.

An example: at Alpin Sun, the biggest worry came from the field teams who enter the data. There was no reliable internet on site, so how would the numbers reach the office? The answer was tablets for data entry and Starlink satellite internet on site. Objections like this belong in the plan before anything is built, not after.

Checklist for your first automation

  1. Collect three to five recurring tasks from the team
  2. Work out the yearly cost of each
  3. Score them with the table above and choose the pilot
  4. Measure the baseline: time, turnaround, errors
  5. Build the pilot in two to six weeks and run old and new side by side
  6. Measure the result and move on to the next candidate

Common questions

Do we need our own IT department?

No. For a first step you mainly need the person who does the task today and someone who can make decisions. An external partner can handle the build and, if you want, the ongoing support.

Do we have to replace our software?

Usually not. Good automation connects the programs you already use instead of replacing them. At Alpin Sun, the existing Excel workbook stayed at the centre.

Will this cost jobs?

The goal is relief, not cuts: the time saved goes into customers, quality and growth that there was no capacity for before. Tell your team that clearly.

Not sure where to start?

That is exactly what our automation audit is for: we analyse your processes and deliver the three to five automations with the biggest payoff, with effort and savings estimates. The first step is a free 30-minute call where we look at your candidate together. You can see more of our work in our portfolio.

Have a process that should run by itself?

Book a free call