Choosing the first task to software-development">automate matters more than how it is automated, and most businesses choose badly.
They pick the biggest, most visible process, because it feels like the one worth the effort. That process is usually complicated, full of exceptions, and touches six people, which is why it stalls.
Four tests for a good first task
It happens often. Daily is ideal. Something done twice a year is not worth automating no matter how annoying it is.
It follows rules. If the answer is always derivable from the inputs, a machine can do it. If it needs judgement about a particular customer, it cannot, and pretending otherwise produces something people override and then stop using.
It is boring. The person doing it will welcome losing it, which means you will get help rather than resistance.
And the result is checkable. You can tell immediately whether it was done right. That matters because the first automated thing has to earn trust, and trust comes from being visibly correct for a few weeks.
What this usually points to
Not the big system everybody imagines. It points to things like the bill being raised the moment a delivery is marked complete, the acknowledgement going out the moment an enquiry arrives, or the reminder appearing on the right person's screen on the right day.
Small, frequent, rule based, checkable. Each one removes a few minutes many times, which adds up faster than one large project that is still being argued about in month four.
And it builds the appetite for the next one, which is the real reason to start small.
Apply the four tests to your list. Which task passes all four?
Choosing the right first task, and deliberately not starting with the biggest one, is how we keep these projects from stalling. Tell us your list and we will tell you which one we would start with.
Comments
Nothing here yet. Be the first to say something.