Where to start automation in a small business: four thresholds instead of general analysis

Four thresholds instead of one big analysis

Automation guides usually start the same way: "first, analyze the company's processes." In practice, that sentence ends the topic for weeks. No one knows where such an analysis should start. So check one process at a time, using four thresholds.

The first threshold is repeatability. If a process looks identical every time – the same set of steps, the same order – it's suitable for automation. If it requires a different owner decision each time, not yet.

The second threshold is a hard deadline. Settlement in KSeF has a statutory deadline that can't be moved by a single day. A process with such a deadline is a natural candidate, because a mistake in it costs more than a delayed email.

The third threshold is dependency on one person. If a process only works when a specific receptionist is actually at work, that's not a process – it's a single point of failure.

The fourth threshold is cost, counted from the other end. Not "how much does implementation cost" but "how much does this mess cost you every month" – in lost inquiries, in delayed responses, in hours spent retyping by hand. If that monthly loss is larger than the one-off work needed to close it, the threshold has already been crossed and further arithmetic will not change it.

A customer inquiry becomes a task, not a sticky note

Let's take a process that looks the same in many companies: a client writes or calls with a question about an offer. Today, that question lands in an inbox, in Messenger, or on a note by the phone. It often stays there.

The first process to automate has a simple chain:

An inquiry from a form, from WhatsApp, or from Google Business Profile goes straight to CRM, instead of to an employee's private inbox. The system itself creates a task and assigns it to a specific salesperson, with a response deadline. The owner doesn't need to ask what's happening with a given lead – they see it in the report, based on data, not someone's memory.

This is the first form of managing the company based on facts, not a spreadsheet in someone's head – before the rest of the automation comes.

Who and what handles the first process in the company

Behind the first process are specific people, not an abstract "system." The receptionist or salesperson still talks to the client – automation doesn't take that conversation away from them. It takes away what happens after it. An inquiry from a phone, WhatsApp, or Messenger goes to CRM instead of a notepad.

Connections between tools are most often handled by n8n – it receives the inquiry, saves it in CRM, and inserts a deadline in Google Calendar for the salesperson. If the company uses wFirma or enova365 for invoicing, the automation only reads data from them. It doesn't replace the accountant.

The technical side – accounts, permissions, task order – is handled by an administrator or consultant. Responsibility for the process itself, i.e., whether the client gets a response on time, stays where it was: with the salesperson and the company owner.

What breaks when a process is manual

A manual process breaks quietly. There's no alarm when the salesperson forgets to call back – there's only a client who, after a few days, writes to the competition. The error is visible only in the sales report at the end of the month. By then, it's too late to change anything.

The most common scenario: the receptionist goes on sick leave, and no one else knows which inquiries are waiting for a response. Without automated inquiry handling, that knowledge sits only in her inbox and in her head. When the process goes into the system, the inquiry doesn't disappear along with the person who just didn't come to work.

This is fixed by an administrator or consultant who sets rules – who gets the task if the first person doesn't respond within an hour. This isn't about better discipline in the team. It's about whether the process has any backup at all.

How much the first automation costs

The cost depends on which threshold you've just crossed, not on the general "project scope." Query automation is the first chain:

If it's only about connecting the CRM with the rest of the tools, that is a different and shorter stretch of the same road – which is why it costs differently.

Reports come next, and they close this chain. A report showing the owner how many inquiries turned into tasks and how many got a response on time is already the first step toward analytics showing numbers, not gut feelings – without it there is no telling whether the earlier links changed anything. If the company also wants to qualify leads automatically, that work runs every day and counts as maintenance, not as an implementation with an end date.

There is nothing here to add up, because the configuration starts from the process you choose first, not from a list. We'd rather work it out together with you, after a conversation about where the chain breaks today. We don't promise a specific number of recovered inquiries or a specific ROI timeline – honestly, no one can calculate that in advance.

When it's not worth starting with automation

Automation doesn't make sense where every situation is different. If a salesperson negotiates a different discount and different term each time, there's no repeatable process to save in the system. There's a conversation that must be conducted by a human.

The same applies to a company that doesn't yet know what its sales process looks like. Before you automate anything, that process must exist in a repeatable form. First, a human does it, several times in a row, the same way. Automation consolidates a process that already works – it doesn't create it from scratch.

It's also not worth starting when the team doesn't have time to learn a new way of working in the next few weeks. Launching requires someone – receptionist, salesperson, accountant – to spend an hour or two checking whether the system does what it was supposed to do. Without this, control over the process stays on paper, not in practice.

Questions we hear most often

Do you need CRM to start?

No. You can start recording the process even in a simple CRM set up for this one case. What's more important is that inquiries no longer land in employees' private inboxes. The price of CRM automation is worked out after a conversation, and the specific tool depends on what the company has today.

How long does it take to launch the first process?

It depends on the number of systems to connect – one form and one CRM is a different scale of work than CRM, calendar, and invoicing together. We don't give a deadline upfront because it also depends on how much data needs to be organized first.

What if we already use wFirma or enova365?

Automation doesn't replace those systems, it just uses them. A tool like n8n pulls data from them and passes it on, for example to a report for the owner. The accountant still works in their system – they don't have to move from it.

What if the process changes every time?

Then it's not yet a candidate for automation – see the section on when it's not worth starting. The process must first stabilize, preferably after a few weeks of doing it the same way. Only then is it worth coming back to the topic.

Let's talk about which process in your company has already crossed the first threshold.