The question we get asked most often isn't "how does automation work" but: "Where do we start?"

And it's the right question. Because many companies get the first choice wrong: they pick the wrong process — too complex, too full of exceptions, too dependent on human judgment. Here's how to avoid that.

The golden rule: high impact, low complexity

The first process to automate isn't the most important one in the company. It's the one that gives you the best ratio between value gained and difficulty of implementation.

  • High impact + low complexity = start here
  • High impact + high complexity = do it later, once you have experience
  • Low impact + low complexity = optional
  • Low impact + high complexity = don't do it

The first project should be an easy win. Not because it's trivial, but because it needs to prove its value before you invest in bigger things.

The 5 questions to ask yourself

1. How often does it happen?

A task that repeats 50 times a day is a much better candidate than one that happens twice a month.

2. How predictable is it?

If the process follows clear rules and has few exceptions, it's easy to automate. If every case is different, it's not the right process to start with.

3. How many people does it involve?

The more people touch a process, the more time is lost on coordination and waiting. Automating a flow that involves 3-4 people frees up time for everyone.

4. How much does it cost today?

How many hours a week does it take up? How many errors does it cause? If you can't estimate it, you probably don't have enough visibility into the process — and that's already a problem.

5. How frustrating is it for the team?

If the team hates a task, automating it immediately builds motivation and trust in the project. It's the best way to overcome resistance to change.

Good first processes

In our experience, these work best as a first project:

  • Email/request sorting — high volume, clear rules, immediate impact (see the customer assistant demo)
  • Data entry between systems — repetitive, error-prone, hated by everyone
  • Generating recurring documents — quotes, reports, standard communications (see the demo from site inspection to quote)
  • Follow-ups and payment reminders — simple rules, big time savings (see the receivables and deadlines demo)
  • Repetitive checks and verifications — missing fields, inconsistent data, attachments

Notice what they have in common: they're all frequent, predictable, and low value-add for the person doing them.

What NOT to choose as your first process

  • The company's most critical process. If it fails, the damage is huge. It's not the place to learn.
  • Processes full of exceptions. If "every case is different," it's not the right time.
  • Tasks that require creativity. Automation handles rules, not ambiguity.
  • Processes nobody has ever mapped. Map first, then automate.

Why the first project matters so much

Your first automation project isn't just an operational improvement. It's a precedent.

If it goes well, the team trusts it. Management sees the value. The way opens for the next project. If it goes badly, "automation doesn't work" becomes the mantra for the next two years.

That's why we insist: don't pick the most ambitious project. Pick the one that wins.