How to identify which processes to automate in your company

To work out which processes to automate in a company, and above all which one first, you need to assess each candidate against two criteria: the impact automating it would have (hours, errors, response time, cost) and the effort it requires. With those two data points, every process falls into one of four quadrants, and the first automation should come from the high-impact, low-effort one. If you're not yet clear on what automating involves, we recommend starting with our guide to business process automation.

In this article we explain how to build the list of candidates, how to score them and how to use the impact-effort matrix to decide based on evidence, not gut feeling.

In 30 seconds:

• Not everything is worth automating: starting with the wrong process wears the team out and delays results.
• A good candidate is repetitive, frequent, rule-based and currently causes errors or delays.
• The impact-effort matrix sorts processes into four quadrants: Quick wins, Strategic, Optional and Discard.
• The first automation should be a quick win; strategic ones come later.

---

Why choosing the first process well matters

The first automation sets the tone for everything that follows. If it delivers visible results, the team adopts it, management trusts it and it becomes easier to tackle more complex processes. If you pick a process that is too ambitious, infrequent or poorly defined, the project drags on, the return never arrives and the conclusion is usually "this automation thing doesn't work here".

That's why, before thinking about tools, it pays to spend time deciding what to automate.

Step 1: list your candidate processes

Get the heads of each area together and ask them about the tasks that take the most time or cause the most problems. Some useful questions:

• Which tasks are repeated every day or every week, always in the same way?
• Where is information copied from one tool to another?
• Which processes depend on a specific person being available?
• Where do errors appear that later need fixing?
• What delays or bottlenecks do customers notice?

It's common to come out of that conversation with a longer list than expected. If you're short of ideas, our article with 25 examples of business process automation has candidates by department.

Step 2: check whether each process can be automated

For each process on the list, look for these five signs:

1. Repetitive: it is done the same way, or almost the same way, every time.
2. Frequent: it happens daily or several times a week.
3. Rule-based: it can be explained as "if this happens, do that".
4. Based on digital data: the information lives in emails, documents or tools, not just in someone's head.
5. With a measurable problem: it costs hours, errors, delays or money.

The more signs it ticks, the better the candidate. If a process relies heavily on personal judgement or changes every time, perhaps only part of it can be automated (for example, data gathering), leaving the decision to a person.

Step 3: score impact and effort

Now it's time to put rough numbers on it. For each candidate, rate from 1 to 5:

Impact, based on:

• Hours spent on the process per year.
• Errors and rework it causes.
• Effect on the customer (response time, complaints).
• Risk of depending on a single person.

Effort, based on:

• Number of tools involved and whether they can be connected.
• Number of exceptions and special cases.
• Quality of the starting data.
• Changes it requires from the team.

To estimate the hours, use a simple formula: people × hours per week × 52. You can do it with the ROI calculator.

Calculate the return on automation

Step 4: place each process in the impact-effort matrix

With both scores, each process falls into one of these four quadrants:

• Quick wins (high impact, low effort). These are the candidates for your first automation. They deliver visible results with little complexity. Typical examples: entering orders from email into the ERP, logging leads in the CRM, payment reminders.
• Strategic (high impact, high effort). They are very valuable but require connecting several systems, cleaning up data or changing the way people work. They are planned after the first ones, building on what has been learnt. Example: integrating CRM, ERP and invoicing end to end.
• Optional (low impact, low effort). Simple improvements that can be made when there is capacity, but they shouldn't distract from the priorities. Example: an automatic internal alert that saves a few minutes.
• Discard (low impact, high effort). They aren't worth it, at least for now. Example: automating a process that happens a few times a year and has lots of exceptions.

The rule is simple: start with the quick wins, plan the strategic ones, leave the optional ones for spare capacity and park the discards.

Before and after prioritising

This is how the approach to automation changes when you prioritise with a method:

Illustrative prioritisation example

Let's look at an illustrative example (not a real case) with four processes at a distribution company, using an hourly cost of €25:

• Entering orders received by email: 2 people × 5 hours/week × 52 = 520 hours per year, about €13,000 per year. High impact, low effort. Quadrant: Quick wins.
• Full CRM-ERP-invoicing integration: affects several teams and a lot of data. High impact, high effort. Quadrant: Strategic.
• Internal alert when a large order arrives: 1 person × 0.5 hours/week × 52 = 26 hours per year, about €650. Low impact, low effort. Quadrant: Optional.
• Annual supplier report with different formats every year: few hours per year and many exceptions. Low impact, high effort. Quadrant: Discard.

In this example, the first automation would be order entry. Once it's up and running, the experience gained and the cleaner data make it easier to tackle the strategic integration.

How it fits into the working method

This prioritisation is at the heart of the first two phases of our methodology:

1. 01 Assessment: processes are identified and their starting point is measured.
2. 02 Prioritised proposal: they are placed in the matrix and ranked by return.
3. 03 First automation: the most profitable one is implemented and validated with real data.
4. 04 Rollout: it is extended to other processes and the team is trained.
5. 05 Continuous improvement: indicators are reviewed and the next step is proposed.

If you want to see which processes are usually prioritised in each area, visit the solutions by department hub. And to translate priority into euros, see how to calculate the ROI of automation.

Common prioritisation mistakes

• Choosing out of enthusiasm, not data. The most eye-catching process isn't always the one with the highest return.
• Ignoring effort. A high-impact process with messy data may be strategic, not quick.
• Automating a broken process. If the manual process isn't clear, it needs simplifying first.
• Not revisiting the matrix. What is strategic today may become quick once systems are already connected.

Frequently asked questions

Which processes should be automated first?

Those that are repetitive, frequent, rule-based and currently cause errors or delays, provided the effort to automate them is low. They are the ones the matrix places in the quick wins quadrant.

What do I do with strategic processes?

Plan them for after the first automation. They require more effort but tend to be the most valuable in the medium term. It's best to tackle them once the team has seen results and the data is in better shape.

How do I estimate impact if I don't have exact data?

A reasonable estimate is enough to get started: how many people do the process, how many hours a week and their hourly cost. What matters is measuring it the same way before and after.

Who should take part in prioritisation?

The heads of the areas involved, someone from management who can make decisions and, if there is one, the IT or systems manager. Whoever does the process every day usually knows its exceptions best.

Can just part of a process be automated?

Yes, and it's often the most sensible option. For example, automating data gathering and validation and leaving final approval to a person.

---

We'll help you decide where to start

Does this process sound familiar? We'll help you measure what it costs you. In the assessment we review 2 or 3 processes with you, place them in the impact-effort matrix and tell you, with numbers, which one to automate first.

Request your process assessment