You signed up to run a mission, or a kitchen, or a craft. The software was supposed to stay invisible: useful, quiet, someone else's problem. Instead the tool catalog comes to you now, through newsletters and conference talks and the board member with a favorite app, and every item in it is demonstrated by somebody having a wonderful month.

Underneath the noise sits the honest question, the one that can feel unsophisticated in a board meeting: does my business need AI at all? It deserves a better answer than a shrug or a sales call. The answer is findable in an afternoon, on paper, for free.

The tool catalog is organized by capability. Your operation is organized by decisions. Start where your week is already organized.

Four questions, one afternoon

1. What keeps eating the week? Skip the interesting problem and find the recurring one: the task that shows up every week and takes a person with it. The order-ahead list. The report that runs late. The numbers somebody has to pull by hand before any decision can happen.

2. Which decision waits on a signal? If one number arrived on Monday instead of Thursday, what would you do differently by Tuesday? If the honest answer is "nothing," write that down too, because it is an answer.

3. Where does the data already live? The strongest AI work connects systems you already run instead of replacing them. At Oliver's, the Chicago restaurant in our case studies, the week's demand signal came out of Toast, OpenTable, a weather feed, and a hotel-occupancy tracker. No new software was bought to get it.

4. Who keeps the habit alive? Every running tool needs a person who checks it, feeds it, and notices when it drifts. If that person exists on your team, good. If the answer is nobody, the first project needs to be smaller, or it needs to wait.

The honest answer is often not yet

A tool adopted because everyone else has one is a chore with a login. It needs attention in exactly the weeks when your mission needs more of you, and it gives back what unmaintained tools give back: nothing, slowly.

Not yet is a respectable answer. The tools will still be there when a real problem shows up, and they will be cheaper and better. What matters is knowing which situation you are in, and the four questions above are how you find out without buying anything.

What passing looks like

Both engagements in the case studies passed this test before anything got built. The restaurant's nagging problem was staffing against unknown demand, the signals were already there, and the decision that got better was planning the week instead of reacting to the day. The research nonprofit's nagging problem was a slow, specialized search, and the method it uses now reruns as the science moves. A small problem with a real signal behind it, and a person to keep the habit: that is a passing score.

And if a pitch is already on your calendar, the four questions in The chatbot nobody asked for will take it apart in ten minutes.

Run it this week

Four questions, on paper, before the next vendor call: what keeps eating the week, which decision waits on a signal, where the data lives, and who keeps the habit. If the answers hold, you will know exactly what to ask for, and a paid discovery can map the first piece of it. If they wobble, you have saved yourself a workshop.

Most teams that did not sign up to run software do not need to. Yet.