# Where is your team with AI?

Eight questions on rules, measurement, agents and costs. It takes a few minutes. You see the result right away, with no sign-up and no email.

## Question 1 of 8: How many of your developers work with AI tools at least once a week?

0. Almost nobody, or we don’t know
1. A few, each on their own
2. Most of them, each in their own way
3. Almost everyone, with shared practices

## Question 2 of 8: Do people have clear rules on which AI tools they may use, and with which data?

0. We have no rules
1. Only a ban or a vague recommendation
2. Written rules and approved tools
3. Rules, approved tools and company accounts for everyone

## Question 3 of 8: Do you know which of your AI uses fall under the EU AI Act and what follows from it?

0. We haven’t looked into it
1. We know it applies, but have no overview
2. We have a list of our AI uses and their risk categories
3. We have the list and an owner, and meet the obligations, such as AI literacy

## Question 4 of 8: Do you measure what AI changes in your software delivery?

0. We don’t measure it
1. Only by feel or with a survey
2. We track metrics, but without a baseline
3. We have a baseline from before AI and compare regularly

## Question 5 of 8: How do your people learn to work with AI?

0. Everyone on their own
1. We share tips now and then
2. The team went through a training
3. Training, internal champions and shared practices right in the repositories

## Question 6 of 8: Do AI agents do work on their own, such as opening pull requests or fixing bugs from the backlog?

0. No
1. Some of us try it
2. Regularly on selected tasks, reviewed by a person
3. As a normal part of delivery, with rules, review and measurement

## Question 7 of 8: Do you know what AI costs you, and for what?

0. We don’t know
1. Only the total invoice
2. By team or by tool
3. By team and by model, compared with the benefit

## Question 8 of 8: How do you control which systems and data AI agents can reach, for example through MCP servers?

0. We don’t; everyone connects whatever they like
1. We know about it, but have no overview
2. We have an approved list of servers and permissions
3. We manage access centrally, with an audit of who did what and when

[See my result](/en/ai-adoption-check)

---

- https://prokopsimek.cz/en/ai-adoption-check
- [Česky](https://prokopsimek.cz/ai-adopce-test)