Most automation projects that disappoint started with a tool. Someone saw a demo, bought a licence, and then went looking for something to point it at.
We start the other way round, with a spreadsheet and a few blunt questions.
Do the maths first
For each repetitive task, write down how often it happens, how long it takes each time, and how many people do it. Multiply it out to get hours per month.
Say your team captures 1,400 supplier invoices a month at four minutes each. That's over 90 hours a month, more than half of one person's working time, spent typing numbers from PDFs into the accounting system. That's a candidate.
A task that takes two hours once a quarter isn't, however much people complain about it.
Look for rules, not judgement
Good candidates follow rules you could write on a whiteboard. If the invoice matches a purchase order and the amounts agree, post it. If it doesn't, send it to Thandi in finance.
If the honest answer to "how do you decide?" is "it depends", automation will struggle. AI can help at the edges, reading a messy email or a scanned form, but someone still needs to check its work.
Check the data before anyone builds anything
If one supplier is spelled four different ways across your systems, cleaning that up is job number one. It's dull, and skipping it is one of the most common reasons an automation falls apart in its first month.
Start where you can count the result
Choose something where before and after are easy to measure: hours spent, response time, error rate. You'll want those numbers when you ask for budget for the next project.
Then build the smallest version that works. One mailbox and one document type is plenty for a first go.
If you'd like a second pair of eyes on your list, that's what our Automation Audit is for.