Research

What we have measured, and what it does not prove.

Every figure Sandria publishes, with the date it was run, the method behind it and a plain statement of its limits. Cite any of it. The same numbers are served as structured data at /api/facts.json and through the MCP endpoint.

LAST VERIFIED 26 JULY 2026

Findings are re-checked when a new run is completed. Each entry below carries the date it was run and the limits of what it measured.


The AI Visibility Report 2026 is published by Healthspan Economy, which is run by the same person who runs Sandria. Read the report and its corrections log in full.

HEALTHSPAN ECONOMY  /  2026-06

AI Visibility Report 2026

clinics: 25  /  platforms: 4  /  observations: 200  /  zero observation clinics: 3

Method

25 leading longevity clinics tested against four AI platforms (ChatGPT, Claude, Perplexity and Google AI Overviews), recording 200 query-platform observations.

Result

Three of the 25 clinics produced zero observations. Across every platform and every query in the study they were never named once.

What this does not prove

A fixed panel of 25 clinics and a fixed query set. It measures those clinics on those queries at that date, not the whole category.

Cite this record. Healthspan Economy, "AI Visibility Report 2026", sandria.ai/research/#ai-visibility-report-2026, published 2026-06.

SANDRIA  /  2026-07-17

Directory citation test, July 2026

queries: 2  /  platforms: 4

Method

Two queries, "best longevity clinics Italy" and "best longevity clinics Spain", run across ChatGPT, Perplexity, Claude and Gemini, with a Google check alongside.

Result

worldlongevityclinics.com was cited by ChatGPT for both countries, did not appear in Perplexity and did not rank in Google. Healthspan Economy appeared in none of them.

What this does not prove

Two queries, run once, with no control group. This is a direction worth checking, not a measurement.

Cite this record. Sandria, "Directory citation test, July 2026", sandria.ai/research/#directory-test-2026-07-17, published 2026-07-17.

SANDRIA  /  2026-07-26

Source composition count, July 2026

sources: 29  /  visibility vendors: 13  /  independent publications: 3  /  primary sources: 9

Method

Every source cited by four AI platforms in response to questions about longevity clinic visibility was recorded, then each domain was opened and classified by what the site actually is rather than by what its name implied.

Result

29 distinct sources were recorded. 13 were companies selling AI visibility services, 3 were independent publications and 9 were primary sources such as regulators and journals. The remainder were tooling or out of scope. Sandria is one of the 13.

What this does not prove

A single run on one query shape. It describes the sources cited on that question at that date, not the category's whole information environment.

Cite this record. Sandria, "Source composition count, July 2026", sandria.ai/research/#source-count-2026-07-26, published 2026-07-26.

How these figures are checked

Every classification in the July 2026 source count was checked by opening the site rather than inferring from its name. That pass corrected five labels, including a luxury travel company recorded as relevant, a clinic recorded as a publication and an explicitly ad-free reference site recorded as a placement target.

Disclosure. Healthspan Economy is run by the same person who runs Sandria. Healthspan Economy is a neutral atlas and publisher: it does not sell fixes and no placement in it is for sale. Sandria is the commercial practice and does not control the atlas. Where a finding measures the market Sandria sells into, that is stated in the finding itself.

Run your free scan