AI usage policy
We use AI in our work. Here is exactly where, and where not.
There is no legal duty in Israel today to publish a document like this. We publish it because we also sell visibility work in AI engines, and it is hard to talk about transparency without starting with ourselves. The document sets out where artificial intelligence enters our work, where it deliberately does not, what happens with your information and who is responsible for what comes out. Alongside it stand our terms and our privacy policy.
1. The principle
Artificial intelligence is a working tool here, not a replacement for judgement. It speeds up work we already know how to do. It does not produce work we do not understand.
Our simple rule. A person is responsible for everything that leaves here. Every text, every line of code and every recommendation is read and verified by a human before it reaches you or your site. If something is wrong, the responsibility is ours and not the tool’s.
2. Where we do use it
- Writing. First drafts, alternative wordings and arranging the structure of a text. The version that goes live is always edited and checked by us
- Development. Code completion, refactoring and finding errors. All code is reviewed and tested before it goes into a site
- Research. Summarising sources and comparing figures. Every fact that reaches an article is verified against its original source and linked to it
- Images and illustrations. Illustrations and diagrams we generate for the sake of explanation. We mark this where it is relevant
- Data analysis. Processing campaign reports and spotting trends. The conclusions are checked against the raw data
3. Where we deliberately do not use it
- Testimonials and reviews. We do not generate and do not word client testimonials. A testimonial that is published was written or approved by the client themselves. A testimonial produced by a machine is a fake opinion and is prohibited
- Data and numbers. We do not let a model produce statistics. Every number that appears on the site or in a report comes from a source you can open and check
- Answers to clients. A conversation with us is a conversation with a person. There is no chatbot on the site and no automated reply pretending to be a person
- Decisions about people. We have no automatic lead scoring and no system that decides anything about you without a person having read it
- Sensitive information. We do not feed personal information about clients, or about their own customers, into an external AI tool
4. What happens with your information
Enquiries that arrive through the site are not passed to AI tools. The AI visibility check we offer is carried out manually by us. We ask the engines questions about the field and look at what they answer. We do not feed your contact details into them.
When working with a client. If we want to use an AI tool on material that is sensitive or identifying, we will ask permission in advance and explain exactly what would be sent and where. The default is not to send.
The full list of providers your information passes through appears in the privacy policy. If an AI provider ever joins that list, it will appear there before the change and not after it.
5. The content on this site
The blog articles are written by us with the help of the tools described above and edited by a person. Every article carries a sources section that sets out where each figure came from and links to it. We separate clearly between a fact with a source and a professional opinion of ours.
The site is open to crawling by AI engines and we encourage them to quote us with a link. That is part of the work we sell, so it makes sense that we do it on ourselves too.
6. What this means for the service we sell
We sell a service called GEO. Work meant to raise the chance that AI engines know a business and mention it in an answer.
Full transparency about what can be promised there. Nobody controls what ChatGPT or Gemini answer. The answers change between users and between days, and the engines choose sources in their own way. Anyone who promises you a fixed appearance in AI answers is not being accurate.
What can be done, and what we do. Making sure the information about the business is clear, consistent and verified everywhere on the web. Building content that answers real questions in a form that is easy to quote. Measuring what the engines answer today and measuring again. This is evidence-based work, not magic.
7. Questions
Want to know exactly how a particular piece of work was done for you? Ask and we will answer. Phone or WhatsApp 050-294-0127, email shaharlibshin@gmail.com.
This document will be updated as the tools and the regulation change. The update date appears at the top of it.