Scope a tool that does one job well
One user, one task, one measure of success, before anything gets built.
The business of AI
A simple way to decide whether a task needs a new tool at all.
5 min readPlain English
In one sentence
Before you buy or build an AI tool, check whether the task is worth changing at all and whether the assistant you already have can do it.
Think of it like…
It’s like sorting out a cluttered storeroom. Sometimes a cleanup is enough, sometimes you buy ready-made shelving, and only now and then do you pay a carpenter to build something to measure.
Run the task through four checks, in order. Many tasks stop at the first or second.
Whichever you choose, decide up front how you’ll judge it and who approves its output.
Good at
Bad at
Skip: the answer was a prompt. A plumbing firm wanted a tool for job follow-ups. Saved instructions in the assistant it already pays for did the job.
Buy: a busy shared inbox. An online shop trials a help-desk tool with AI drafting built in, on last month’s real emails, before signing up.
Build: quotes only you can price. A print shop’s quotes follow its own rules, so it links a form, a spreadsheet and an assistant, and a person approves every quote.
Myth vs reality
Myth: “We need a custom AI built for our business.”
Reality: Many small-business tasks are handled well by the assistant you already have, given clear instructions and your own examples. Build only when a task is frequent, specific to you and worth maintaining.
Your team writes the same kind of follow-up email every day. What should you try first?
One user, one task, one measure of success, before anything gets built.
Every inquiry lands in the right place with the key facts pulled out. A person assigns it.
Three kinds of help, and how to tell which one a vendor is selling you.