Methodology
How the check calculates
A measuring instrument whose calculation cannot be inspected is not one. This page discloses the complete evaluation scheme — the same rules the browser uses to compute the result.
What is measured?
The check measures an organization's readiness for AI and agents across six dimensions: Leadership, People, Technology, Agent, Innovation, and Future Readiness. Three questions per dimension establish the current state; each question has five substantively ordered answer levels from 0 to 4. The order of answers in the interface follows the substantive order — point values are deliberately never displayed.
How are the scores computed?
Each dimension sums its three questions to 0–12 points; the total ranges from 0 to 72. A dimension counts as “AI-ready” from 8 of 12 points — this normative threshold is always drawn in the radar. The total maps to four maturity levels: watching (0–17), experimenting (18–35), systematic (36–53), transforming (54–72).
What is the bottleneck?
The bottleneck is the weakest of the six dimensions — on a tie, the first in schema order. The recommended action at the end of the check comes from crossing bottleneck and type: 24 texts plus a dedicated text for the Refuser.
What are pathology markers?
Fifteen answer levels carry markers — typical pathologies such as Shadow Use, Ban Reflex, Head Monopoly, or AI Theater. When certain markers occur together, they condense into six combination findings (for example Underground, from Ban Reflex plus Shadow Use). Markers do not affect the scores; they sharpen the diagnosis.
How is the type derived?
Two axes condense the six dimensions: ambition (Leadership, Innovation, Future — does the organization want it, and does it know what for?) and substance (People, Technology, Agent — can it deliver?). Both axes are normalized to percent; the threshold sits at 50%. This yields four types: Pioneer, Visionary, Engine Room, Spectator. If an axis lies within 8 percentage points of the threshold, the result is flagged as borderline.
What is the Refuser?
A special case outside the four quadrants: if the total is below 22 points and at least two of the three markers Downplaying, Ban Reflex, and Immunity Myth are set, the organization has decided — explicitly or not — against engaging at all. That is not a maturity level; it is a state before the journey.
What data is stored?
On completion, the browser transmits anonymous numbers: the 18 answer levels, the values computed from them, optionally industry and size class, and the completion time. No IP address, no name, no email, no cookies. Runs under 60 seconds are flagged as fast clicks and excluded from all evaluations. Comparison values (average, top quartile) appear only from 100 data sets per cell — anything earlier would be noise. The result itself never depends on the server: it is computed entirely in the browser and lives as a link.
The six dimensions
| # | Dimension | Axis | Range |
|---|---|---|---|
| 1 | Leadership Readiness | Ambition | 0–12 |
| 2 | People Readiness | Substance | 0–12 |
| 3 | Technology Readiness | Substance | 0–12 |
| 4 | Agent Readiness | Substance | 0–12 |
| 5 | Innovation Readiness | Ambition | 0–12 |
| 6 | Future Readiness | Ambition | 0–12 |
The four maturity levels
| Points | Level | Meaning |
|---|---|---|
| 0–17 | watching | AI is happening — elsewhere. First observations, no structure. |
| 18–35 | experimenting | Individual trials and tools, still without direction or breadth. |
| 36–53 | systematic | Strategy, rules and capability building interlock. |
| 54–72 | transforming | AI and agents are changing processes and the business model — steered, measured, intended. |