AI readiness assessment
The assessment is a short engagement at a fixed fee, agreed before it starts. We study your operation as it actually runs, put a cost on the tasks that repeat, and finish with a priced, ordered plan for automating them. The plan is written to stand alone: you can carry it out with us, with another firm, or with your own people.
The fit
None of these mean your business is badly run. They mean it has grown past the point where memory and effort can hold everything together, which is the situation this engagement exists for.
The method
The work runs in the open, on your premises or your screens, and every finding traces back to something we watched happen. Here is what each part involves.
We map your operation through conversations and direct observation, not questionnaires. Forms get answered with the official version of a process; the real version includes the workaround someone invented three years ago and the step everyone does twice to be safe. So we ask the people doing the job to show us, in the real screens, while we take notes. The gap between what the manual says and what actually happens is usually where the cost sits.
For each recurring task we establish how often it happens, how long it takes, who does it, and what that person's time is worth. We also count the quieter costs: errors introduced by retyping, and interruptions that pull skilled people off the work customers pay for. The result is a figure per task, in hours and in money, so the plan can rank fixes by what they recover rather than by what is fashionable to build.
Every task gets sorted into one of two piles. If a task can be written down completely — when this arrives, do that — it belongs to conventional software, which is cheaper, faster, and never has an off day. Only tasks that genuinely require reading a situation and weighing it are candidates for AI. Expect the first pile to be bigger than you thought — and that is good news: the boring pile is the inexpensive one.
Where a task does call for judgment, AI does not get the job automatically. Our standard is that an AI component must first pass tests assembled from your own past records — real inquiries, real documents, real edge cases — before it touches anything live, and any output a customer will see gets a human sign-off first. The plan marks which tasks clear that bar today and which should stay with people. When an AI step later gets one wrong, that example is added to the tests it has to keep passing.
The build list is sequenced by how quickly each item earns back its own cost. Each entry carries a price, the hours it should recover, and what has to exist before it can be built. Ordering by payback means early items can fund later ones, and it gives you a natural stopping point: build until the returns thin out, then stop. Nothing obliges you to do the whole list.
The plan is written so it does not depend on us. Each item is described twice: in plain language for you, and in enough technical detail that any competent developer could build it. Tools and services are named with their running costs, and nothing in it is proprietary to our firm. Hiring us for the build is one option, and deliberately not the only one.
The shape
A first conversation settles which parts of the business the assessment covers. We quote one fixed fee for that scope, in writing, and nothing starts until you have agreed to it. If the scope grows later, the fee changes only with your sign-off.
We talk with the people who carry the daily load and watch the work happen — on site where practical, over screen sharing where not. Your team's part is a handful of scheduled conversations; nobody fills in a form. Any system access is scoped to what observation requires and agreed with you first.
We sort the findings on our own time: tasks classified, costs attached, candidate fixes priced and put in order. If something we saw does not add up, we come back with questions rather than guesses.
We walk you through the finished plan, item by item, and answer whatever it raises. Then it is yours. There is no obligation to build anything, with us or at all — and if the honest finding is that little is worth building, the plan states that plainly.
The deliverables
Everything below arrives as documents stored in your accounts, not ours. They are the deliverable; the engagement is complete when you hold them.
Next step
If you are weighing an assessment, start with a short call. We will ask about your week, tell you what we would examine first, and give you an honest read on whether the fee is likely to earn itself back.
Set up a callAlso in the catalog