Insights
No filler. Just useful answers about search and AI visibility.
Tests, research, and working models designed to help you make a better decision, not keep you scrolling.
GetBorvis / Organic visibilityEach guide answers one specific question, shows where the evidence came from, and explains what is worth doing next. No recycled listicles. No invented ranking secrets.
Is GEO worth it for a small business? Put the price next to the evidence.
Retainers start around $1,500 a month. Across 45 reviewed studies no technique shows stable cross-platform effects. We sell this work and still say most should wait.
Read the guide ↗A redesign can pass visual QA and still erase the pages search engines already trust.
The handoff register we use before a website launches, one row per old URL, with binary acceptance tests, a rollback trigger and the AI-era fields older migration checklists do not have.
Read the guide ↗18 of 48 agency sites published llms.txt. Sixteen days later, 20 did.
We checked the same 48 endpoints twice. Adoption moved, and 23 returned HTTP 200 while only 20 passed the format screen. Both aggregate datasets published.
Read the guide ↗Search Console now reports generative AI impressions. It does not report prompts.
Google rolled the report out worldwide on 31 August 2026. Four dimensions are documented and query is not one of them, which changes what you are allowed to claim from it.
Read the guide ↗GEO, AEO and SEO. The definitions keep moving, and there is a reason for that.
A survey of 45 studies found terminology and evidence standards remain heterogeneous, and no technique with stable cross-platform effects. The labels cannot settle while the measurement underneath them has not.
Read the guide ↗Earned media drives 84% of AI citations. That is not the argument for digital PR it looks like.
Journalism is only 27% of that figure, and the top cited domain per engine is Wikipedia, PubMed Central and Reddit. None of the three is something you can pitch.
Read the guide ↗Google says topical authority is not a thing. Its own documentation is more interesting.
We checked the leaked Content Warehouse attributes against the primary source, found the field locations reported across SEO are wrong, and read them beside an active Google patent that undercuts how topical authority is sold.
Read the guide ↗Schema markup does not get you cited in AI answers.
A controlled test on 1,885 pages, Google saying it in plain language, and 300,000 domains all point the same way. Including where we had it filed wrong.
Read the guide ↗Google AI and ChatGPT do not play by the same rules.
What each system relies on, what overlaps, and why one visibility strategy cannot measure both properly.
Read the guide ↗How AI decides which businesses to name.
What actually happens between a buyer asking a question and a company being recommended, straight from the platforms’ own documentation.
Read the guide ↗Do brand mentions matter more than backlinks for AI visibility?
The 75,000-brand study is real and it excluded every domain below Domain Rating 40. Four signals, measured separately, with the confounder nobody controls for.
Read the guide ↗If a claim sounds certain, the evidence should be visible.
Platform documentation is treated as fact. Vendor research is labelled as research. Observed patterns are presented as interpretation, not dressed up as secret ranking factors.
Why GetBorvis works this way ↗