Research
We measure how AI assistants answer buying questions in the UAE and publish what we find, including the unflattering numbers. Stated methods, dated figures, and vendor claims taken only from vendors' own pages.
Reports and guides
We measured 2,061 AI citations across six engines. Vendor websites supplied 5% of them.
Six engines answering 22 software buying questions cited 25 third-party hosts 2,061 times, and the vendor's own site 106. Reddit alone was cited more than any roundup site. ChatGPT cited three outside hosts in three weeks; Perplexity cited 23. The full source anatomy, with the cited pages opened and read.
UAE local businesses are ranking on Google and disappearing from AI answers
A car-servicing directory scoring 62 out of 100 on traditional search scores 3 out of 100 in AI answers. We measured two UAE sites daily across six engines and published the scores, the method and the mechanism behind the gap.
From SEO to AEO: what actually transfers, what does not, and why
Ranking well and being cited are different outcomes with different inputs. Which of your existing search work carries over, which is wasted effort, and why Search Console is the only dataset that connects the two channels.
AI visibility for small businesses: what it costs, and what you can safely ignore
Published entry prices run from $29 to $499 a month, and most of that is built for teams with a dedicated analyst. What an owner-operated business actually needs, what it can skip, and a free ten-minute self-check.
Building an AEO service line without making it unprofitable
Easy to sell, easy to lose money on. Platform selection on delivery economics, engagement scoping, why you price the fix rather than the dashboard, and the four traps that turn this into unbilled work.
How we publish research
- Every number is traceable. Measurements come from stated methods on stated dates. Vendor claims come from vendor pricing pages with retrieval dates. If we cannot source it, it does not go in.
- We publish unflattering scores. The sites in our reports are ones we have full measurement access to. We anonymise them, but we publish their real numbers, including the bad ones.
- We do not score what we have not tested. No composite indices, no accuracy ratings, no rankings we cannot reproduce. Where we have not done the work, we say so.
- We disclose commercial interests. Where a page recommends a tool we have a financial interest in, it says so next to the recommendation and links to our disclosures.
- We correct in public. Found an error? Tell us and we will fix it and note what changed. Vendors can request corrections to their own rows.