Knowledge base
Who stands behind this
The knowledge articles are published under the company's name, not under an individual's. This page explains that decision, who reviews the articles, and how you can hold us to it.
Publisher, not author
The knowledge articles are published by Cybervize Consulting GmbH. No individual author is named, and that is deliberate: these articles are a company offering, not a collection of personal essays. Nobody should be personally on the record for 274 texts whose legal basis keeps moving.
We know a named author would help discoverability. Search engines and AI answer services weigh authorship. We considered that and decided against it. What would otherwise be missing here is carried by what follows: a named editorial team, a review path you can follow, and limits stated openly.
The review path
Every article passes the same stations before it appears.
- Drafting against the sourceThe article is written against the primary source: standard, legal act, circular or catalogue. Where the version matters it travels with the reference, for instance “MaRisk circular 06/2024 (BA)”. Where a framework is named without an edition, please check the version in force yourself. Whether an article says so explicitly varies.
- CountersignatureA subject-matter owner countersigns the article. Only then does it appear. The signature is logged in the corpus. The article page does not show it: the log mixes several kinds of event, and we would rather give no date than a wrong one.
- Re-signature on changeWhen the source changes, the article is updated and signed again. A purely technical redeployment does not count as a review.
Article pages currently show no countersignature date. The corpus log keeps gate run, countersignature, source status and re-check as separate events, and for some signatures the date lives only in the signing record. Any single number taken from it would be wrong for a share of the articles. The “Stand” note at the foot of an article means something else: the time of the last import from the corpus, not of a review.
What the articles deliberately do not do
Each article carries rules that define the permitted depth of answer. They apply to the text on this site as much as to the assistant that cites from the same knowledge base.
- Standards are not quoted verbatim. We reference and explain; we do not replace the original text.
- No article reproduces a standard clause by clause, nor offers a checklist meant to stand in for the original.
- Drafts are labelled as drafts. A consultation paper is not law in force, and we do not present it as such.
- Where a statement rests on the absence of a public source, we say so instead of asserting a gap.
These rules are not a marketing promise: the field exists per article and is filled for most of them. For eight of 274 the list is empty, and the general rule applies there. They bind the editorial team as much as the assistant. What a separate check step before publication verifies is only that the internal rule fields themselves do not leak into the text; whether a paragraph goes too far is decided by the countersignature, not by a script.
Sources and limits
Every article names the frameworks it relies on. For some the version is stated, for others only the framework name: there the article does not give an edition, and you should check it yourself before use. What binds in every case are the originals issued by the responsible bodies. Our articles classify and translate into practice; they replace neither the standard nor legal advice in an individual case.
Found an error?
If you spot an outdated version, a wrong reference or a misleading classification, write to us. A correction on the substance is worth more to us than an unchallenged archive.
Report a correction