What we automate first in a company that has never automated anything

You automate the process that already has a clear input, a clear output, and a person who is sick of it.

Not the one that looks good in a board deck.

Not the one the software vendor told you would "transform the customer experience."

You start with the spreadsheet that makes Sarah groan every Tuesday morning.

Here is why.

Most companies that have never automated anything pick the wrong first project. They choose something that sounds impressive. A chatbot to handle customer queries. An AI that predicts inventory. A dashboard that pulls data from five systems and generates insights. The project gets budget, a timeline, a kickoff meeting with a consultant. Then it stalls.

The reason is simple. Those projects are fuzzy. The input is not one thing. It is a stream of messy human conversation, or a dozen file formats, or a business rule that changes depending on who you ask. The output is even fuzzier. What does success look like? A happy customer? A stockout prevented? Nobody can measure it cleanly. So the project drags on for months. The person doing the work learns that automation means more meetings, not less work. Sarah keeps copying invoice data by hand. When someone mentions AI in the break room, she rolls her eyes.

That is the cost of picking the wrong first step. Not just the money or the time. The cost is Sarah's trust. The cost is the quiet signal that spreads through the company: automation is just another overpromised initiative that does not change anything on the ground.

The automation that changes someone's week

Now flip the choice.

You look for the process that already has sharp edges. A single, predictable input. A single, predictable output. And a person who dreads it.

That process lives inside your company right now. It is the daily report someone compiles from three CSV files and emails to the ops team. It is the customer inquiry that arrives in a standard form and needs to be classified, tagged, and routed to the right rep. It is the invoice data that lands in a PDF in a shared mailbox and has to be retyped into the ERP. The person doing it knows every step by heart. They could do it in their sleep. And they are sick of it.

That person is your first automation customer. Not the executive who wants a shiny AI demo. Not the investor who wants to see innovation. The person whose Tuesday morning gets lighter.

When you automate that process, three things happen fast.

First, the result is visible. Sarah opens her laptop and sees the invoice data already entered. She does not need to check a dashboard. She just sees the work done.

Second, the feedback is instant. If something breaks, Sarah will tell you within minutes. She knows the process better than anyone. She will help you fix it because she wants it to work.

Third, the trust compounds. Sarah tells the two colleagues next to her. One of them has a similar grind. They ask if you can automate theirs. That is how automation spreads in a company that has never automated anything. Not through a company wide announcement. Through a quiet Tuesday morning when someone's repetitive work simply disappears.

How to find it in under ten minutes

You do not need a process mapping workshop. You do not need an external consultant to spend a week interviewing people. You need one question.

Walk over to a team lead, or send them a short message. Ask: "Which task makes someone groan every time it comes up?"

Not "which task takes the most time" or "which task is most expensive." Those questions lead back to the fuzzy projects. They lead to the inventory prediction model that will take six months to build and another six to tune. The groan question leads you straight to the sharp edged process. The one with a clear input and output. The one someone is already sick of.

Once you have a few answers, look for the process that meets three conditions.

The input arrives in a consistent digital format. A spreadsheet, an email with a predictable subject line, a form submission, a PDF that always has the same structure.
The output is a single, well defined action. A row added to a database. A file saved to a folder. A notification sent to a specific channel.
The person who does it can explain the rules in plain language. "If the invoice amount is over $1,000, I flag it for review. If the customer type is wholesale, I apply a different template."

That is your first automation candidate. It does not need to involve AI. A simple script or a no code flow can handle it. But if the rules are messy, or the input varies in a way that simple logic cannot catch, that is where AI helps. The point is not the tool. The point is the process.

Start with that process. Build the automation this week. Run it for a few days with the person watching. Tweak it. Then let it run on its own.

What happens next Monday

Next Monday, that person opens their laptop. The work is already done. Or it is half done, waiting for a quick human check. Either way, their groan is gone.

That is the only metric that matters for your first automation. Not ROI. Not efficiency gains. A person who used to dread a task no longer dreads it.

From there, you have something no slide deck can give you. You have an internal reference. A real example that someone in the company can point to and say: "They did it for Sarah's invoices. They can do it for my order confirmations."

You can build that first flow yourself with no code tools and a few afternoons of work. Or you can have a team handle it for you, from process mapping to monitoring, for a fixed monthly fee. Nexibeo does exactly this for companies that have never automated anything. One or two processes a month, built and maintained, so the person who groans gets relief without you managing a project.

Either way, start with the groan. Start with the sharp edges. Your first automation is not a strategy document. It is a Tuesday morning without Sarah retyping invoices.