Automation is not free. It costs money to build and attention to maintain. So before we quote, we run the same arithmetic with every client.
Step one: measure the task honestly. Minutes per run, times runs per month. People consistently underestimate this - so measure for a week rather than guessing.
Step two: cost the time. Take the fully-loaded hourly cost of whoever does it, not their salary divided by hours.
Step three: add the error cost. What does a mistake in this task cost you? Missed follow-ups, wrong invoices and stockouts all carry a real number.
Step four: compare against build plus annual maintenance. Assume maintenance is roughly 15-20% of build cost per year.
If payback lands under twelve months, build it. Twelve to twenty-four months, only if the task is also error-prone or blocks growth. Beyond twenty-four months, leave it alone and automate something better first.
The common mistake is automating the most annoying task instead of the most expensive one. Those are frequently different tasks.
