Research

We show our working.

Most AI visibility statistics you’ll read come from asking one engine one question, once, with no list of what was asked. Here’s how we do it instead.

Research we’ve conducted

Every piece carries the exact questions we asked, the market, the engines and how many times we ran it.

  1. Piece 2

    Ask three AI engines who's best in Dubai: the overlap is low, not zero

    Across seven Dubai categories, ChatGPT, Perplexity and Gemini named 147 businesses and 10 appeared on all three. The same engine asked twice also returned different lists below first place.

    Read the research

  2. Piece 1

    The Dubai companies writing the articles AI quotes about them

    A 44-answer test in Dubai found companies' own 'best of' articles still ground Perplexity's answers but not ChatGPT's, and that self-published list positions are lost to competitors running the same play.

    Read the research

  3. Piece 3

    Gemini isn't answering your question. It's routing you into Google.

    A 44-answer test of ChatGPT, Perplexity, Gemini and Google Search in Dubai found Gemini routes answers into Google's own feeds, which feed depends on the category, and Google's AI Overview already covers four of seven categories.

    Read the research

  4. Report formatPlaceholder

    Placeholder: how a category scan is reported

    Placeholder scaffolding showing the research template — tables, citations, method and limitations. No findings. Delete this file to remove it.

    Read the research

  5. Method notesUpdated 19 August 2026Placeholder

    Placeholder: method notes and citation practice

    Second placeholder, showing how method, sources and update history are recorded. No findings. Delete this file to remove it.

    Read the research

Your category, measured the same way

Same method, same disclosure: the questions, the market, the run count, and where every mention came from. No charge, and nobody will chase you afterwards.

Show me where I stand