AI implementation and AI consulting are not the same purchase.
One ends with a document. The other ends with something running. Knowing which you are buying is most of the decision.
Consulting tells you what to do. Implementation does it. Most of the market sells the first and calls it the second, which is why owners end up with a roadmap and no change to their week.
The shapes are different
Consulting
Usually sold in stages. An assessment of what you run, then a roadmap of what could change, then a build quoted separately. What you own at the end of the first two stages is a document.
Right when the question is genuinely open and the decision is expensive.
Implementation
Sold as the thing working. Whatever counting has to happen, happens as part of doing it, and what you own at the end is work that runs without anyone starting it.
Right when you already know the work is heavy and you want it off your week.
One question tells you which you are being sold
Ask what you will have on the last day. If the answer describes a document, a strategy, a readiness score or a roadmap, that is consulting, whatever the page calls it. If the answer describes something running that was not running before, that is implementation.
Neither answer is wrong. Paying for the first while expecting the second is the mistake.
Where the money usually goes
The common pattern is that the counting stage is priced on its own, the plan is priced on its own, and the build is priced again after both. A business that already knows which part of its week is heavy is paying for two stages of finding out.
If you want to know the size of the thing before you talk to anyone, work out the hours yourself. It takes a minute and it costs nothing.
Questions people ask.
Is one better than the other?
No. They answer different questions. Consulting answers what you should do. Implementation answers who is going to do it. A business that already knows what it wants is paying twice if it buys the first.
Can I get the assessment without the build?
From most consultancies, yes, and that is the normal shape: an assessment, then a roadmap, then a separate build. The thing to check is whether you are left holding a document you now have to find someone to act on.
What if I do not know what I need yet?
Then some form of the first conversation is genuinely useful. What matters is whether it ends with a document or with something running.
Skip the assessment stage.
Thirty minutes on what you actually run. If the answer is that you do not need us, we will say so.
Book a walkthrough