Most businesses don’t have a software problem.

They have processes that grew organically.

Someone exports an Excel file. Someone else cleans it up. Another person copies the data into a different system. Then someone sends an email asking for approval.

Nobody designed it this way. It just happened.

The good news is that these processes are often excellent candidates for automation.

You can find them by asking a few simple questions.

1. Where does work wait?

Look for processes where something sits in someone’s inbox, spreadsheet, or queue waiting for the next person.

Ask:

• Where do we wait for someone to respond?
• Where do approvals slow things down?
• What regularly takes days even though the actual work takes minutes?

Example:

A customer submits a request.

An employee reads the email, checks the CRM, looks up information in another system, prepares a response, and waits for a manager to approve it.

The actual work might take 10 minutes.

The process takes two days.

That gap is an automation opportunity.

---

2. Where are people copying information?

Manual data entry is one of the easiest places to look.

Ask:

• Where do we copy information from one system to another?
• Are people downloading CSVs and uploading them somewhere else?
• Does the same customer, order, invoice, or product get entered multiple times?

Example:

Website form → Excel → CRM → accounting system

If a human is acting as the integration between your systems, software might be able to do that job.

---

3. Where are spreadsheets holding the business together?

Spreadsheets aren’t bad.

But when a spreadsheet becomes a database + workflow engine + reporting system, things start getting interesting.

Ask:

• Which spreadsheets does the team depend on every day?
• What happens when the person who maintains one is away?
• Are people constantly filtering, cleaning, merging, or exporting data?
• Does someone have to run a spreadsheet manually before another process can start?

Example:

Every Monday someone spends three hours downloading sales data, cleaning it, calculating numbers, and creating a report.

That isn’t really a reporting process.

It’s a manual data pipeline disguised as a spreadsheet.

---

4. Where does the same work happen repeatedly?

Repetition is a strong automation signal.

Ask:

• What do we do dozens or hundreds of times every week?
• Which tasks follow almost exactly the same steps?
• What does a new employee have to learn that could be a button?

Example:

For every new customer, someone:

1. Creates a CRM record
2. Creates a folder
3. Sends a welcome email
4. Creates an internal task
5. Adds the customer to a spreadsheet
6. Notifies accounting

If the steps are predictable, there is a good chance software can orchestrate them.

---

5. Where do mistakes and rework happen?

Errors are often more expensive than the original work.

Ask:

• Where do people frequently make mistakes?
• What has to be checked twice?
• What gets sent back because something is missing?
• Where do we regularly correct data?

Example:

An employee receives an order, manually enters 15 fields, and another employee checks them.

You might not need a better checker.

You might need to remove the manual entry.

---

6. Where are humans acting like routers?

This one is easy to miss.

Sometimes an employee’s main job is simply moving information between people and systems.

“When X happens, send this to Sarah.”

“When the customer does Y, update the spreadsheet.”

“When the amount is above €5,000, ask John.”

These are often workflows waiting to become software.

Ask:

Could we describe the decision as a few clear rules?

If yes, part of the process may be automated.

---

7. What is the bottleneck?

Don’t automatically automate the biggest or most annoying process.

Find the part that actually limits the business.

Ask:

• Where does work pile up?
• Which person or team is always overloaded?
• What prevents us from processing more customers/orders/projects?
• If we doubled demand tomorrow, what would break first?

Example:

Your sales team can generate 200 leads a week.

But someone manually qualifies every lead and can only process 50.

The bottleneck isn’t lead generation.

It’s qualification.

Automating that step may have a much bigger impact than automating something else.

---

One final question

After you’ve found a repetitive or slow process, ask:

What would happen if we didn’t have to do this manually anymore?

Don’t start with technology.

Start with the outcome.

Maybe you need:

• a small automation between existing tools
• a better internal dashboard
• an approval workflow
• automated data processing
• an integration between two systems
• a custom internal application
• or a SaaS product because the process has become part of your actual business model

The technology comes afterwards.

First find the process that shouldn’t require a human every time.

If you can identify a process that is repetitive, slow, error-prone, or dependent on one person, it is probably worth investigating.

And you don’t need to automate the entire business.

Start with the one process that keeps getting in everyone’s way.