# The 10-point AI skills check

> Ten questions on where your AI skills live, who owns them, and whether they survive the person who knows. Two minutes, scored in your browser.

Source: https://www.skilder.ai/en/ai-skills-check/

## The ten questions

Each question is yes or no. A yes scores one point.

1. Do we have one central place where approved AI prompts and skills are stored?
2. Can a new hire access our proven AI workflows on day one, without asking a colleague?
3. Does every business-critical AI workflow have a named owner?
4. Is there a review step between someone writing a skill and the whole team using it?
5. Do we know which systems and data each AI agent is permitted to access?
6. Can we see who used which AI capability, and when?
7. When a skill produces a wrong answer, is there a defined way to report and fix it?
8. If our most advanced AI user resigned tomorrow, would their workflows survive?
9. Do we version our AI instructions, so we can see what changed and roll back?
10. Are our AI capabilities grouped by role, rather than assigned person by person?

## How to read your score

### 0–3 / 10: Nothing is captured yet.

Your best people have working methods, and none of them would survive their resignation. Start by making one team’s way of working explicit, then reuse it.

### 4–7 / 10: You have habits, not a system.

Parts of this work, unevenly, and nobody can say which parts. Which ones to close first, and in what order, is the conversation worth having.

### 8–10 / 10: You are already governing.

The practices are there. What is missing is leverage: more roles, faster, without the person who knows becoming the bottleneck.

## What closes a gap

What closes each gap depends on context: the missing building block, the order to take them in, and which of them actually matter at a given size. That is what a 30-minute call with the skilder team covers. This document does not reproduce it.

Book a call: https://www.skilder.ai/en/ai-skills-check/