Most of the money businesses waste on AI is not wasted on bad vendors. It is wasted on good tools bought at the wrong time, by a business that was not yet ready to use them. The failure looks like the AI's fault. It almost never is.
So before you sign anything or pay anyone - us included - run these six checks. They take about ten minutes and they will save you far more than that. If you can answer all six, you are ready. If you cannot, the honest next step is cheaper than whatever is being sold to you.
1. Is the thing you want to improve actually written down?#
AI works on information. If the process you want to improve lives on paper or in someone's head, there is nothing for AI to act on yet.
Ask: is the data this project needs already captured digitally? If your stock, sales, or records are not in a system, that is a stage 0 job - and it is cheaper and more valuable to do first. No answer here, and nothing else on this list matters.
2. Can you name the one job, in one sentence?#
"We want to use AI" is not a goal. "We want customer WhatsApp enquiries about stock and opening hours answered automatically" is.
Ask: can I describe the single task I want done, specifically enough that I would know whether it worked? Vague goals produce expensive, unfocused projects. A sharp, narrow job is the one that succeeds.
3. Do you know what it costs you today?#
You cannot judge whether AI is worth it if you do not know the price of the current way. How many hours a week does this task eat? What do the mistakes cost?
Ask: what is this problem costing me now, in time or money? If you cannot put a rough number on it, you will not be able to tell whether the AI paid off - and you may be automating something that was not worth the effort.
4. Have your people used AI at all yet?#
A team that has never used AI as a simple assistant will struggle to work alongside one embedded in their process. Stage 1 fluency is what lets people brief the system and judge its output.
Ask: does my team already use AI for everyday tasks? If not, that is the step to take first - cheap, fast, and it builds the exact skill a bigger project depends on.
5. Is it clear what happens when the AI is wrong?#
Every AI is sometimes wrong. A system you can trust is not one that never errs - it is one where the errors are caught and handled. Before you buy, you should know where the human checkpoint is.
Ask: when the AI makes a mistake, who catches it, and how? If the answer is "it just won't," walk away. A serious partner designs for the wrong answer, not just the right one.
6. Are you buying the right rung?#
Look at where you honestly sit on the ladder, and at what is being sold to you. If you are at stage 1 and someone is selling you a stage 4 "predictive analytics platform," the mismatch is the problem, no matter how good the product is.
Ask: is this the next rung above where I actually stand - or two rungs too high? The right purchase is almost always the step just above you.
What a "no" means#
If you answered no to any of these, that is not bad news. It is the most valuable thing this checklist can give you: a reason not to spend yet, and a clear, cheaper step to take first. Every "no" points to an earlier rung - and climbing it will make the eventual AI project work instead of fail.
Where to go from here#
If you ran the six checks and you are ready, the next move is a straight conversation about the one job you named in check two. If you are not ready yet, the checks told you exactly which rung to climb first.
Either way, we would rather tell you the honest answer than sell you the wrong one. Tell us what you run, and we will tell you plainly where AI would help and where it would not - and if the answer is "not yet," we will say so. For the full map, see the five stages of AI adoption.