On this page
What is business process automation?
Business process automation is turning work that is done by hand, over and over, into flows that run on their own. The aim isn’t to take anyone’s job away; it’s to win back the hours people spend copying and pasting.
In most businesses this work is invisible. Nobody writes “I spend six hours a week merging reports” in a status update, because over time it has simply become part of the job. When the process is mapped, the picture usually surprises everyone. It’s also the application area that comes up most often in my AI consulting work.
Which tasks suit automation?
There’s a practical way to tell whether a task is worth automating. If the answer to these five questions is yes, it’s a strong candidate.
- Does the task repeat every week, always in the same way?
- Can its steps be written down, so one person could explain it and hand it to someone else?
- Does the input come from a defined place and the output go to a defined place?
- Is it free of judgement calls, i.e. are the rules clear?
- If it goes wrong, can the mistake be reversed?
The fifth question is the one most often skipped. Where an error can’t be undone, you can still automate, but a human approval step must sit in the middle. We decide this at the design stage.
I go through these five questions with examples in a separate article. For a quick first assessment of one of your own tasks, try the automation calculator.
The flows I automate most often
| Flow | Before | After |
|---|---|---|
| Reporting | Data downloaded from several dashboards and typed into a spreadsheet | Data is collected automatically and the report is ready on the agreed day |
| Incoming requests | Forms, emails and messages scattered; some get missed | Every request lands in one place, is categorised and has a draft first reply waiting |
| Content production | Every piece starts from scratch; reformatting done by hand | One source produces several formats; the approval step stays with a person |
| Transfer between systems | Copy and paste from one tool to another | Records sync automatically; an alert is sent if something fails |
| Documents and proposals | The template is filled in by hand each time | Data drops into the template automatically; the final check stays with a person |
Notice that in every row the final decision stays with a person. That is deliberate. Automations without an approval step run fast, but when something goes wrong it takes a long time to notice.
The automation process
| Stage | What happens | Typical duration |
|---|---|---|
| Process mapping | Laying out, step by step, how the work actually runs | 3–5 days |
| Pilot selection | Choosing one flow with high impact and low risk | 1 day |
| Build | Building the flow and running it on test data | 1–2 weeks |
| Live test | Running it alongside the manual process for a while and comparing | 1 week |
| Handover | Training, documentation, what to do if it fails | 2–3 days |
Skipping the fourth stage is risky. Running the new flow in parallel with the old method for a while is the only reliable way to compare results and build trust. A week of patience prevents months of doubt.
How I choose tools
I don’t sell any particular automation platform. Depending on the need, that might be visual automation tools such as n8n, Make or Zapier, language models, your existing software’s own integrations, or custom code where necessary.
I look at three things when choosing: your team must be able to understand and maintain it, it must not lock you in as your needs grow, and its monthly cost must be proportionate to the time it saves. If those three aren’t met, we look for something simpler. Where off-the-shelf tools fall short, custom AI solutions come into play.
Where does automation fail?
This section is here for a reason: a build that ignores the common failure modes will repeat them.
- When there’s no written process. Automation speeds up a process that already exists. Automating a vague process only speeds up the vagueness.
- When only one person understands it. A system only one person understands stops when that person goes on holiday. That’s why documentation is part of the delivery.
- When failure isn’t planned for. If the flow stops one day, who finds out? Automation without alerts breaks silently and you notice weeks later.
- When the gain isn’t measured. If you didn’t measure how long the task took before, you can’t prove what you saved afterwards. We set up measurement from the start.
- When the team is left out. If the person doing the work by hand today is excluded from the build, the system won’t be adopted. They know the process better than anyone.
What you receive
- A working automation, built on your own accounts
- A process map and a document explaining what was automated and why
- A short guide on what to do if something fails
- Handover training for the team
- Optional maintenance and improvements at agreed intervals
Frequently asked questions
Will automation take my employees’ jobs?
In practice it’s usually the opposite. The work I automate is the repetitive, draining work nobody wants to do. The team spends those hours talking to customers, checking quality and making decisions. I also don’t recommend making automation decisions without the team.
We’re a small business. Is it too early for us?
Repetition matters more than size. A three-person team might spend eight hours a week merging reports, and for that team those hours are worth more than for a large company. For most small businesses the sensible start is one flow, then building from there.
Will it work with our existing software?
Most modern software allows outside connections. During discovery we list the tools you use and I check how each can be connected. If a system can’t be connected, I say so at the start rather than trying to work around it.
What happens if it breaks after handover?
Every flow gets failure notifications, so nothing stays silent when it stops working. The documentation includes the first things to check. If you want support beyond that, we can arrange ongoing maintenance; it’s optional.
Do you build with n8n, Make or Zapier?
Yes, depending on the need these platforms can be used. We choose based on whether your team can maintain it, your data requirements (n8n, for example, can also run on your own server) and whether the monthly cost is proportionate to the time saved. We clarify the process first and choose the platform afterwards.
How much time will it save?
There’s no honest answer before the process is measured, and I’d be wary of anyone who promises a number. What I do is this: we measure how many hours the task really takes before the build, and repeat the same measurement afterwards. If there’s a gain you’ll see it; if there isn’t, you’ll see that too.
Related articles
Which Repetitive Tasks Should You Automate? A Five-Question Filter
Not every repetitive task is worth automating. A practical five-question filter, a worked example and a prioritisation table for deciding which work…
How to Write an AI Usage Policy for Your Team
If your team has started using AI tools, you need a written rule. The seven sections of a one-page policy people actually…
How we start
The starting point is always the same: a free 20-minute discovery call. I don’t try to sell you anything in that call. We talk about how the work runs, where it gets stuck and what is worth automating.
The call ends in one of two ways. Either we find an area where working together makes sense and I send you the scope in writing, or I tell you that you don’t need me for this. The second happens often, and it’s better than wasting anyone’s time.
20 minutes · free · online
Choose a time that suits you
- Times are shown in your own time zone (UK or Türkiye)
- No preparation needed; one recurring task is enough
- In Turkish or English
or send an email: fahriaksoy.tr@gmail.com

