Every task you do twice a week is a bug

Most businesses do not have an obvious automation problem. Nobody is manually entering ten thousand rows a day into a spreadsheet — that gets fixed immediately, because the pain is loud. The automation that actually gets missed is quieter: a task small enough to just do, done often enough that it adds up to real hours nobody is tracking against anything.

The test we actually use

If you catch yourself doing the same manual task twice in one week, write it down. Do not automate it yet — just note it. After a month, look at the list. Anything that shows up most weeks is a candidate. Anything that shows up every week, without exception, is not a candidate — it is already costing you real time and you have just not priced it yet.

Pricing whether it is worth fixing

The maths is simple and worth doing on paper, not in your head, because "it only takes a few minutes" almost always undercounts. Take the real time per instance — including the context-switch back into whatever you were doing before, which is real and rarely counted — multiply by how many times a week it happens, multiply by roughly 48 working weeks a year, and compare that annual hour figure against what it would cost to have someone (or something) do it instead. A task that takes four minutes, five times a week, is roughly sixteen hours a year. That sounds small until you price sixteen hours at whatever your time is actually worth, and compare it against the cost of a workflow that removes it permanently.

Why this category gets ignored

Big, painful problems get fixed because they are impossible to ignore. Small recurring ones survive because each individual instance is genuinely not worth stopping to fix — the five minutes it takes to do the task by hand is less than the time it would take to figure out how to automate it, in the moment. That is true for any single instance and false for the whole year of instances, and almost nobody does the second calculation until someone points out the total.

What we build first on every automation project

When a client says "automate the business," the honest first step is not the most impressive workflow available — it is a short audit of exactly this list: what gets done by hand, more than once a week, that has a clear trigger and a clear output. Those are the workflows with the fastest, most measurable payback, and they are almost always less glamorous than the automation the client originally asked for. The flashy one usually gets built too — just second, once the boring one is already saving hours every week.