Process Automation: Where Should Your Business Start?

Process automation starts with choosing the right process, not with buying software. In most businesses the same data is typed into 3 programs, approvals are chased on WhatsApp, and the monthly report is rebuilt by hand. This first part of a 7-part series on process automation explains in plain words what automating a process means, gives you a scoring method with 4 criteria to choose your first process, and an exercise on 10 repetitive tasks you can run yourself this week. If you want the general case for automation before the method, start with Business Process Automation to Cut Manual Work.

Three familiar scenes

Before any definitions, look at an ordinary day in a mid-size business:

• The same data in more than one place. A customer order arrives on WhatsApp, an employee types it into an Excel sheet, retypes it into the sales program, and a colleague types it again into the warehouse program. 3 entries for one piece of information, and every entry is a fresh chance of error.
• An approval chased through chats. A purchase or leave request is sent as a photo in a WhatsApp group and waits until the manager sees it; when it is late, nobody knows where it stopped.
• A report built from scratch every month. Someone spends days copying figures from several files and formatting them, then the manager finds a number that does not match the others. We covered this exact scene in The Reports You Used to Build by Hand Every Month.

Nobody is falling short in any of these scenes. People do the work with the tools they have, and they know better than anyone where the process hurts. The problem is that human effort goes into moving and repeating instead of judging and following up.

What automating a process means

A workflow is the sequence of steps a request passes through from start to finish. Automating it means the system takes over the repetitive steps according to clear rules. Every automation has 4 parts:

• The trigger: the event that starts the process, such as a request arriving, a document being uploaded or a date being reached.
• The rules: the conditions that set the path, for example: if the amount exceeds a set limit, the request goes to the finance manager.
• The action: what the system does, such as creating a record, sending a notification or moving data from one program to another through an integration.
• The exception: the case the rules do not cover, which goes to a responsible person to decide. This part is what makes automation safe.

Automation does not mean removing human judgement, it does not mean cutting staff, and it does not mean buying a tool and then looking for something to do with it. It means people stop repeating what can repeat itself, so they are free for what needs a decision. If your question is about AI specifically, we gathered practical examples in 7 Processes AI Can Run in Your Company Today, and a later part returns to AI reading documents with a human review.

How to choose your first process: 4 criteria

The most common mistake is to start with the biggest process, or the one that annoys the manager most. It is better to score every candidate on four criteria, giving each a score from 1 to 3:

• Frequency: how often does it happen? Once a month scores 1; daily or several times a day scores 3.
• Number of hands: how many people or programs does the request pass through? The more handoffs, the more places to wait and to err.
• Clarity of rules: can the decision at each step be written as a clear if-then sentence? A process that relies on personal judgement every time scores 1 and is not a good start.
• Cost of an error: what happens if a mistake or delay occurs? A wrong shipment, an angry customer or a missed approval scores 3.

Add the scores, so the total falls between 4 and 12. The process with the highest total and the clearest rules is your first candidate. Notice that we give no expected savings here: the real number is the one you measure by timing the process today and again after it is automated. Any general figure you read does not know your business.

The Origami view

We are a Saudi technology company that builds systems, apps, integrations between systems and AI solutions. In any automation project we start from the people doing the work today, not from a screen: we sit with them, score the processes and pick one small process with clear rules, because the success of the first process is what builds the team's trust in the next one.

The automation you need may be a simple link between two programs you already have, or a system built around the process; the choice comes after the scoring, not before. That is the order the rest of this series follows, and what we build is on our services page.

Exercise: score 10 repetitive tasks

Run this exercise over one week, with no new tool:

• Collect 10 repetitive tasks from department heads and from the employees themselves, such as leave and purchase requests, onboarding a new customer, delivery notes, supplier documents, stock transfers between branches, and report preparation.
• Write one line per task: what starts it, who touches it, and where it ends.
• Score each task on the four criteria from 1 to 3, and add them up.
• Time the top 3 tasks yourself: how many minutes one run takes and how many times it repeats per week. That is the baseline you will compare against later.
• Ask the person who does the task: which step bothers you most? The answer often reveals the first place to automate.

If a task handles data about people, such as customers or employees, note it beside the task. The Personal Data Protection Law published by SDAIA requires the controller, in Article 19, to take the organizational, administrative and technical measures and means necessary to protect personal data, including when it is transferred, and any automation that moves data between your programs is a transfer. We cover this with Articles 29 and 31 in the part on control. This is not legal advice; confirm the details with SDAIA or a lawyer.

In the next part

You have picked your first process, but do not automate it yet. Automating a messy process makes the mess faster. Part 2 of the series, Map the Process Before You Automate It, shows how to draw the process as it really happens with the people who run it: the trigger, each step, who does it, where it waits and every exception, how to separate the rule from the habit, and which steps to delete before automating anything.