Is the process ready for automation
Answer about one specific process, not about the company as a whole. Seven questions, two minutes.
How often does this process repeat?
This is about same-type operations: a request processed, a document parsed, a client answered.
What the quiz actually checks
Not a company's readiness for AI — seven questions cannot measure that. What gets checked are the conditions under which automation is capable of delivering a result at all.
There are four of them, and they are the same in any industry.
Repeatability. Infrastructure and support costs barely depend on the number of operations, while the savings depend on it directly. Automation starts to pay back above a certain flow: below that, the arithmetic simply does not add up, and that is a question of volume, not of technology.
The existence of rules. A machine repeats a decision, it does not invent one. If no rules exist and every case is decided on the spot, there is nothing to repeat. Rules may be unwritten — that is fixed by a week of conversations. It is worse when they do not exist at all.
Data availability. If the data is locked inside a system with no access, integration becomes the first project. That gets done too, it is simply a different project with a different budget and a different timeline.
The cost of an error. It determines not «possible or impossible», but the mode of operation: the system decides on its own, or it prepares a decision for a human. The second mode costs more to run and saves less.
What the quiz does not ask
There are deliberately no questions about budget, company size or technical specialists on staff. None of that affects the readiness of the process.
Nor is there a question about management's appetite — although in practice that is what most often decides the fate of a project. But appetite is not measured by a questionnaire, and nobody is ready to admit its absence in a quiz taken about themselves.
What to do with the result
A high score is a reason to count the money and launch a pilot on one section. Not to specify the entire system: a pilot on a small piece shows the real quality of the data, and that is almost always worse than expected.
A middling score is a reason to close the specific weak point that dragged the score down. Usually that is written rules or data access, and both are settled without developers.
A low score is a reason to postpone AI and take on the process itself. Automation works well on top of order and badly in place of it.
FAQ
Why does the cost of an error lower the readiness score?
The more expensive the error, the more human checks have to stay in the process. The system still works, but the savings drop: instead of replacing routine, the result is acceleration with mandatory oversight. That is useful too, it just pays back more slowly.
The rules exist, but they are written nowhere. Is that a problem?
A small one. The rules will have to be pulled out of people's heads and written down — usually that is a week of work with those who run the process. The unpleasant news surfaces in the same place: different employees often hold different versions of the rules, and that has to be settled before automation, not after.
What if the process is rare but very expensive?
Then the maths runs on the cost of a single error, not on volume. Checking a contract before signing, for instance, happens rarely, but a missed clause costs millions. Tasks like that get automated as an assistant to the reviewer, not as a replacement.
The quiz advises starting with the process. What should be done?
Describe the process in words: who does what, in what order and by what rules. Half of all processes straighten themselves out after this exercise, without any AI at all. For the rest, it becomes clear exactly what is missing for automation.
Check it on your case?
We will look at your process and give numbers for it, not for averages.
Related tools
Updated: July 26, 2026