# Five minutes, and you know where you stand.

An AI system answers no better than the knowledge under it. Ten questions measure that knowledge on five points, and then you see where you put Inora to work first.

Ten questions · no email address needed

1. Reachable Whether the knowledge you work from can be found, by people and by systems.
2. Owned Whether every protocol has an owner and a version, and is current.
3. Followed Whether protocols are followed in practice, and whether a deviation is noticed.
4. Traceable Whether a wrong answer can be traced to its cause, and whether you then change the system.
5. Sovereign Whether you have set down where your data and knowledge are processed, that they are never used for training, and that you take your knowledge with you in an open format when you leave.
6. The model Does not count. Your knowledge decides.

Question 1 · Reachable. Your staff need the current protocol in the middle of a shift. How do they find it?

- They ask a colleague who knows.
- They search folders or the intranet, and sometimes find an old version.
- It is in one place, but finding it takes minutes.
- They find it within seconds, on their phones too, and see only what they are allowed to see.

Question 2 · Reachable. Where is the knowledge your organisation works from: protocols, work agreements and handbooks?

- Largely in the heads of experienced colleagues.
- Spread across drives, folders and systems, with no overview.
- In one document system, but not everything is in it.
- In one managed whole, complete and readable by software, with access rights intact.

Question 3 · Owned. Take the ten protocols looked up most often. How many of them have an owner and a version number?

- Nobody knows.
- A few; the rest were written once and never kept up.
- Most, though for some it is unclear who keeps them up to date.
- All ten: each with an owner, a version and a date for its next review.

Question 4 · Owned. A national guideline is revised. What happens to your protocols?

- Nothing, until someone happens to notice.
- Someone updates them when there is time, with no set deadline.
- The owner updates them, but old versions stay in circulation for a while.
- The owner revises them within a set period, and the old version disappears everywhere at once.

Question 5 · Followed. How do you know whether protocols are followed in practice?

- We do not know.
- From incidents and complaints, after the fact.
- From audits and spot checks, a few times a year.
- We see it continuously, and a deviation leads to a decision: the protocol changes, or the practice does.

Question 6 · Followed. Your staff find that a protocol does not fit practice. What do they do?

- They work around it, and nobody hears of it.
- They mention it to a colleague or their manager, and it goes no further.
- They can report it, but rarely hear what happens to it.
- They report it to the owner, who assesses it within a set period and tells them what changes.

Question 7 · Traceable. Someone acts on a wrong answer, from a colleague, a document or a system. Can you find out where that answer came from?

- No, there is no way to find out.
- Sometimes, if someone remembers.
- Usually, after several people have dug into it.
- Always: every answer can be traced to the source and the version it rested on.

Question 8 · Traceable. A mistake has been made. What happens next?

- We find out who did it.
- We sort out the case and move on.
- We look for the cause, but the finding often stays in a report.
- We investigate without blame, write the finding down and change the system, so the mistake does not come back.

Question 9 · Sovereign. What is set down about how your data and knowledge are processed?

- We do not know, or we do not use AI yet.
- We use AI, but we do not know where our data and knowledge are processed or what they are used for.
- The supplier says processing takes place in the EU or in our own environment, without use for training. This is not fully set down in the contract.
- The contract sets down that processing takes place in the EU or in our own environment and that our data and knowledge are never used for training.

Question 10 · Sovereign. Can you run AI wherever you want, on your own servers too?

- We do not know where our data is processed.
- No, only in the supplier’s cloud.
- In a cloud of our choice, but not on our own servers.
- Wherever we want, on our own servers too, and when we leave we take all our knowledge with us.

## Every result starts with one team.

[Talk to our team](https://axiomatic.digital/en/contact#conversation)
