Content Transparency and Use of AI

Who is accountable for the content

Nativ Avivi, founder of NATIV AI, is accountable for the content of this site. He sets the content policy, the rules for choosing topics and the checks an article must pass before publication, and he builds and operates the system that produces the articles.

The byline "By Nativ Avivi" at the top of each article names who is accountable for the article and for the policy it was produced under. It does not mean that every article was written word by word by him.

How a blog article is produced

The blog is produced by an automated process with these stages:

• Choosing a topic: automated analysis of search data, keyword research, top search results and industry updates identifies topics likely to interest business owners.
• Writing: the draft is written in Hebrew by an OpenAI large language model, following NATIV AI's instructions: a direct answer up front, at least two external sources, a practical example and an explanatory diagram.
• Automated checks: before publishing, the system checks minimum length, at least two external links, internal links to existing articles, a valid diagram, and title and description length.
• Publishing: an article that passes the checks is published automatically. An article that fails is held for manual approval by Nativ Avivi.
• Translation: the English version is translated from the Hebrew by a language model and published automatically.

What no person reviews

To avoid any misunderstanding, here is what the process does not include:

• An article that passes the automated checks is published without a person reading it first.
• The automated checks measure structure and form (length, sources, links). They do not verify that every claim is correct, or that every linked source supports what is written next to it.
• The English translation is automatic and gets no separate human review.
• Diagrams are generated automatically from the article's content and are not reviewed separately.

What data the model receives

In the current process the model receives a topic, keywords, public search results, the list of existing articles on the site and style guidelines. It does not receive client information, website enquiries or internal documents.

Limitations: how to read the articles

Language models can make mistakes, state a claim with too much confidence or invent a detail. Requiring external sources is meant to reduce this, but it does not remove the risk.

The articles are meant to explain and give general direction. Do not rely on them as binding legal, financial or technical advice. Before a business decision or an implementation, check the primary source and consult a professional.

Site pages

The text of the site's pages (services, FAQ, About and the author page) is written and edited with the help of AI tools, and NATIV AI is accountable for it.

Found a mistake?

If you find a factual error, a broken link or a misleading sentence, write to us. We will check and correct it, and when an article is corrected, the updated date at its top changes.