The 4 questions to ask before automating anything

Question 1: Is the process repeatable?

If the same input produces the same action 90% of the time, you have something to automate. If it depends on context, judgment, or "it depends on the customer," you have a process that needs a human, with maybe a tool to support them.

Automation is wasted on the unique. Save it for the rote.

Question 2: Is the data clean?

Automation reads from your data. If your data is wrong, the automation will confidently do the wrong thing at scale. Before automating anything, audit the source. Are the names normalized? Are the statuses consistent? Is anyone editing the spreadsheet that the script reads from?

Cleaning data is dull but it's the difference between an automation that helps and one that creates a new full-time job apologizing to customers.

Question 3: Does anyone use the output?

We've built reports nobody opens. We've sent notifications that get muted on day two. Before automating, find the human who needs the output and confirm they actually need it. If the answer is fuzzy, the automation will be too.

Question 4: Will the rules hold for six months?

Some rules are stable for years (invoices on net-30). Some rules change every quarter (this season's pricing). If the rules of your process won't survive a quarter, code is the wrong shape for them. Use a no-code workflow tool, or just keep it manual.

Two no's = don't automate yet. Fix the underlying process first. Then come back.

Sources

1. [1]Internal automation post-mortems, 2024–2025
2. [2]Gartner: Hyperautomation Trends, 2025