AI search / UAE
How to measure AI search visibility responsibly
This is a research protocol, not a published performance study. No UAE visibility index results are claimed until observations have been collected and reviewed.
Define the sample before testing
Specify the sector, geography, company inclusion criteria and prompt set. Publish the selection method and exclusions. A convenient sample must not be described as representative of all UAE businesses.
Keep comparable conditions
Record platform, interface or API, model where disclosed, search mode, language, location settings and UTC timestamp. Use fresh conversations and repeat prompts. Keep different interfaces in separate result groups.
Keep the evidence
Save the full response, cited URLs and an evidence file reference. Record brand mention and own-domain citation separately. Mark failed or unavailable runs as missing rather than a negative result.
Calculate clearly defined measures
Mention rate is valid runs mentioning the brand divided by all valid runs. Own-domain citation rate is valid runs citing its domain divided by all valid runs. Count each run once per measure and report sample sizes and repeat counts.
Review accuracy manually
Check descriptions against current primary sources. Use a second reviewer for ambiguous mentions and keep a corrections log. A model’s confident wording does not verify a business fact.
Publish limits with the findings
Results are a dated snapshot under the recorded conditions. They do not prove a cause, predict all users’ answers or establish that visibility generated revenue. Obtain permission before publishing client-sensitive evidence.
Primary sources
Technical sources reviewed: 20 September 2026. Platform behavior can change.
Make your next step clear
Tell us your website, market and the problem you want to solve.