We sell an AI service, so we are not going to be shy about using AI. We are also not going to pretend a machine wrote nothing. Here is exactly who does what.
SHORT ANSWER
Blog posts on this site start as an AI-assisted draft. Levi Kotka then checks every fact, number and source by hand, rewrites what is wrong, and publishes under his own name. He holds editorial responsibility for every page. No AI edits a page after his review.
What the AI does
- Collects and organises published research, then names the source for every number.
- Writes the first draft of a blog post from that research and from our own notes.
- Builds the page markup, the structured data and the plain-text file for agent crawlers.
- Suggests headings and questions based on what owners actually ask.
The tools used are large language models from Anthropic and from OpenAI, plus web research tools. Nothing is published straight out of a model.
What the human does
- Checks every statistic against the original publication, not against a summary of it. If the original cannot be found, the number is deleted.
- Checks that every source named is real, current, and says what the page claims it says.
- Deletes anything invented. This matters because we have no clients yet, so there are no testimonials, no case studies, no client logos and no results on this site. If you ever see one, it is a mistake, and we want to hear about it.
- Rewrites the language so it reads like a person and makes a plain claim instead of hedging.
- Takes the blame. Levi Kotka is the responsible person for every page here.
What we never publish
- A number without a source.
- A client result, a testimonial or a logo that does not exist.
- A promise of a position inside any AI answer. Nobody can guarantee that.
- A single lucky screenshot of an AI answer presented as proof. AI answers change run to run, so we report shares across repeated runs and show the sample size.
Why there is a label on every post
Article 50 of the EU AI Act (Regulation (EU) 2024/1689) has applied since 2 August 2026. Article 50(4) says that a deployer who publishes AI-generated or AI-manipulated text to inform the public on a matter of public interest has to disclose that the text is AI-generated. The duty does not apply where the text has gone through human review or editorial control by a person who takes responsibility. Article 50(5) says any such disclosure has to be clear, easy to tell apart from the rest of the page, shown at the latest on first exposure, and accessible.
InnoBotZ is established in the EU, and our readers are mostly in the United States. Rather than argue about whether a marketing blog counts as a matter of public interest, we do both things at once: we label every post in plain sight, and the label records the human review that the exception is built on. The label sits directly under the title of every post, not in a menu and not in the terms.
The European Commission published its Guidelines on the scope of Article 50 on 20 July 2026, along with a voluntary Code of Practice on Transparency of AI-generated Content. We follow the substance of both. The hard rule we keep internally: once Levi has reviewed a page, no AI touches it again. A change means the review runs again.
Found a wrong fact? Tell us.
Email info@innobotz.com with the page address and the line you think is wrong. We will check it against the original source. If we got it wrong, we fix the page and note the correction date. No argument, no delay.
Your data
Client data and audit data are not used to train third-party AI models. The full detail is in our privacy policy and cookie policy.