Business automation – where to start when you hear about it everywhere

Five years ago company owners mostly asked about sales, people and margin. Three years ago the question of digitalisation arrived – general, unthought-through. Today a more specific question dominates: where do I start automation. The reason is that automation has stopped being abstract. Every boss has a concrete irritation in mind: the sales rep who spends hours on quotes, the accountant who shuffles invoices from email to the system, the assistant who glues a report together in Excel every week.

The problem, though, is not in identifying the pain, but in moving from pain to action. You open the browser, type business automation, and you see thousands of articles, dozens of tools, hundreds of promises. Everything sounds reasonable, but nothing points to where to put the first step.

This article tries to fix that. It does not sell any tool. It does not promote any methodology. It shows five concrete steps that have worked in hundreds of European companies – small, medium and large. Each of these steps can be done in a week. Each has a clear output. And most importantly, each leaves behind something that actually works – not another deck that the board watches and then returns to its own concerns.

The simplest and most often skipped step. Before you choose a tool, before you hire a consultant, before you email a vendor for a demo, sit down for half an hour. Open a blank page. And write down the three activities in your company that annoy you the most. Not strategic challenges – concrete, daily irritations.

It may be the fact that the sales rep spends two hours a day copying data between the CRM and a spreadsheet. It may be that every Monday at 10:00 someone produces a report read by one person. It may be retyping hours from paper travel claims into the HR system. It may be not knowing whether client X opened yesterday's email.

Collect three such irritations. Then, importantly: ask your team. Most often your list overlaps with theirs at 30%, and the other 70% are things you do not even notice because the team does not raise them. After an hour of conversations you have a list of 10–15 activities, of which 3–5 are obvious automation candidates.

A fuller picture of what European companies most often automate is in our article on what to automate – 10 processes with the highest return.

From the list of 3–5 candidates we pick one. Not all at once. The first automation project has to succeed and convince the organisation it works. The second and third come later, when the first starts to live.

The selection criteria are simple. First: the activity has to repeat regularly. Something that happens once every six months is not a good first candidate. Second: it has a clear output – it is obvious what is supposed to happen. Third: it costs time, not just nerves. Fourth: you would happily automate it yourself, if you could.

In practice one of these usually wins: notifications with concrete signals (new lead, email from Tier A client, vendor invoice above a threshold), automated sending of a template (e.g. order confirmation, follow-up after a meeting), generating the first version of a document (quote, contract, weekly report), electronic forms replacing paper (leave, travel, requests).

The worst first choice: something that requires changing the way 50 people work all at once. The best first choice: something that changes the work of 3–5 people without making them change their daily habits.

This is where most companies make the biggest mistake. They start by comparing ten tools, reading articles about RPA, hyperautomation, AI agents. All of it makes sense – but not at the beginning. The first automation in 9 out of 10 European companies can be built in a tool that is already paid for.

If the company is on Microsoft 365 (most are): you have Power Automate. Standard connectors are included in the Business Standard or Enterprise plan. You can build the first flow in 30 minutes, no developer needed. Power Automate natively connects to Outlook, Teams, SharePoint, Excel, Planner.

If the company is on Google Workspace: you have Google Apps Script and Google Workflows. Less user-friendly than Power Automate, but enough for most first projects.

If the company uses monday.com: you have built-in automations (hundreds of templates), no developer, no consultant. Same for HubSpot, Pipedrive, Salesforce – all have built-in workflows.

Only when no existing tool is enough do we reach for Zapier, Make or n8n (if IT has the skills). But we do not start with them. The choice between automation platforms is covered in our article on Power Automate vs n8n vs Make.

The classic corporate trap: the first automation project turns into a three-month project with a schedule, RACI charts, phases, gates, change management. This is exactly how most first automations die. Because before anything is built, half a year passes and the organisation stops believing.

A better way: a two-week pilot. The first week is building the automation in its simplest possible form. That is: take the chosen activity, open Power Automate (or whatever), and make a working flow in 2–4 hours. It does not have to be perfect. It can have three exceptions. The important thing is that it works.

The second week is releasing it to three, four people. We watch what works, what does not. We note exceptions. We fix. After two weeks either the automation stays and we expand it to the rest of the organisation, or it does not stay and we go back to step one with a new candidate.

In both cases we learnt something at the cost of two weeks, not three months. This is the foundation that no methodology teaches, but every company that really started automating learns: fast, simple, in real life, not on slides.

The first automation works. Three people use it. Those three talk to others, that they no longer have to do this annoying thing. Hunger appears: maybe this, maybe that? This is the decision moment for the whole organisation – to scale chaotically or consciously.

Chaotic scaling: every department builds two or three automations of its own, some duplicate, some conflict, no one knows who built what. A year in, the company has 80 automations of which 50 do not work, but no one knows which.

Conscious scaling: we appoint one person (most often in IT or operations) as automation lead. Every new automation lands in a simple register (one spreadsheet with 5 columns: name, who built it, what it does, when last checked, who it serves). Quarterly review – what works, what does not, what to remove. That costs 4 hours of that person's time per month. Saves dozens of hours of chaos.

The second element of conscious scaling is the candidate queue. The list from step one (3–5 candidates) starts to live. Every month we add new ideas from the team. Every quarter we pick the next one to deliver. After a year the organisation has 8–12 working automations, not 80 chaotic ones.

Conscious scaling is also the moment we stop using only Power Automate and start looking further – Power Apps for forms, Copilot Studio for smarter agents, integrations with other systems. A fuller picture in our article on Power Automate vs n8n vs Make.

Six mistakes we observe in European companies repeatedly. Each can be avoided from day one.

Mistake 1: starting from a big, important process. The board says we are automating all of sales or all of customer service. Six months later it turns out nothing has been automated, because the scale paralyses. Correction: start small, measurable, fast.

Mistake 2: hiring a consultant before picking the first process. A consultant helps, but only when we know what we want. Starting with a consultant gives us a 3-month analysis and a deck, not a working automation. Correction: first contact with a consultant after steps 1–2.

Mistake 3: buying an expensive automation platform. UiPath, Automation Anywhere, Blue Prism – great, but not needed for the start. Correction: start with a tool you already have.

Mistake 4: no owner. Without someone to lead the pilot and own its success, automation dies. Correction: nominate an automation lead at the very start, even part-time.

Mistake 5: automating chaos. Trying to automate a process nobody can describe. Result: the automation inherits that chaos. Correction: if the process cannot be described in 10 minutes, fix it first, then automate.

Mistake 6: no communication with the team. People fear automation means layoffs. Without clear communication the resistance is huge. Correction: from day one say it plainly – automation takes the boring work, not the people. More in our article on automation does not take jobs, it takes chaos.

How much does the first automation cost? In most European companies between EUR 0 and EUR 1k. Zero – if you build it yourself in Power Automate that you already pay for. EUR 1k – if you hire a consultant for 2 days to help with the first pilot. A full automation programme rollout in a 50–250 person organisation is typically EUR 9–22k per year.

How long does the first automation take? One to two weeks, if we walked through steps 1–4 consciously. A week to build, a week to pilot. If it takes longer, it most often means we picked the wrong first activity (too big) or the wrong tool (overpaid for something we did not need).

Who should be the automation lead in the company? Most often someone from IT or operations – but not a developer. The ideal person: someone who understands the business, likes technology and has authority in the organisation. In smaller companies (up to 50 people) this is often the COO or the owner.

Do we need a consultant? Not always. The first 2–3 automations can be built in-house if someone has 20–40 hours to spend across a quarter. A consultant starts to make sense at 5+ automations, integrations between systems and when the organisation wants to scale consciously.

What if the team fears automation? A standard situation. The key is communication from day one. Every automation takes a piece of the boring work, not the person. Show it concretely on the first automation – we saved 30 minutes a day, that person uses those 30 minutes on X, which really matters to the company.

Most companies that started automation consciously share one story. The first automation was small. Concrete. Saved an hour a week for three people. Sounds unnoticeable. But over three months the organisation moved from doubting whether this makes sense to asking what else we can automate. That is the turning point.

The five steps above (notice → choose → choose a tool → pilot → conscious scaling) work in 5-person companies and 500-person companies. The difference is not in scale, but in patience. Most automations die in the first two weeks when someone tries to scale too fast.

A fuller picture of concrete processes to automate is in our article on what to automate – 10 processes. A fuller picture of first steps for a small company is in our article on automation in a small business.