See exactly what the AI sees.
Every number on this page is a real measurement with its instrument and date. Nothing is illustrative. This is the Measure layer; Diagnose, Act and Prove follow, then our own scores.
Five questions its buyers ask, put to four engines. Named in 3 of 20 answers.
| Buyer question | Gemini | ChatGPT | Claude | Perplexity | Named |
|---|---|---|---|---|---|
| Best websites to compare international health insurance in 2026? | no | no | no | no | 0/4 |
| How can I compare IPMI quotes online? | no | no | no | cited, not named | 0/4 |
| Best comparison tools for expat health insurance? | yes | yes | no | no | 2/4 |
| Which sites let me compare Cigna, Allianz and Bupa plans? | no | yes | no | no | 1/4 |
| Cheapest way to find international health insurance as an expat? | no | no | no | no | 0/4 |
Named instead, across the same 20 answers: Pacific Prime 4, International Citizens Insurance 3, Squaremouth 3, ExpatInsure 3, BrokerFish 2, NowCompare 2. "Cited, not named" means the engine linked the page as a source without naming the brand in its answer.
The pages AI quotes and the questions it tests.
For every buying question, the check records which pages each engine cited. These are the sources winning the answers you are absent from: your shortlist of where to earn a mention. Below, the same ipmicompare.com check, 28 Sep 2026.
Pages cited across the 20 answers, by domain
Counts are citations, not answers: one answer can cite several pages from one domain.
Questions tested, and how many of the 4 engines named the site
- 0What are the best websites to compare international health insurance in 2026?
- 0How can I compare IPMI (international private medical insurance) quotes online?
- 2Best comparison tools for expat health insurance?
- 1Which sites let me compare Cigna, Allianz and Bupa international health plans?
- 0Cheapest way to find international health insurance as an expat?
The questions are written from the client's own positioning and agreed before the first check. They stay fixed so that later checks are comparable.
Thirteen checks, fixed weights. Not a vague grade.
Can AI read the site? Thirteen technical checks with weights that never move silently, plus findings that carry no points. A check that cannot be completed (a rate limit, a timeout) is reported as inconclusive, never as a failure. Below, our own homepage.
The full Scout report adds four content checks to these thirteen: one clear heading, FAQ or answer blocks, review or case-study markup, and a clear call to action. That is why the Scout report's AI readiness for this site (90/100, 29 Sep 2026) differs from the website check above (100/100). Two instruments, both shown, never blended.
We do the work. You approve it before it goes live.
The measurement tells us why a site is absent. The work falls into three kinds. Nothing is published, changed or sent without a named person saying yes.
Crawlers refused while browsers are served
The commonest silent blocker: a CDN or firewall rule that returns 403 to GPTBot or PerplexityBot while a browser gets 200. Looks fine to anyone checking by hand. The fix is an edge rule, re-measured the same hour.
One page per buyer question the site loses
Written from the client's own positioning, never a model's guess about it, with the structured data engines extract from. For instlyscout.com: five pages, published 30 Sep 2026, against the five questions in the check above.
Earn a place in the pages engines already quote
Most citations go to independent articles and round-ups, not to vendors' own sites. The source list from the check names them. Pitches are drafted from measured facts; a person sends them.
Before and after, from the same checks, or it did not happen.
A baseline is recorded before any change. The same instrument runs again after. The dated pair is the proof; memory and screenshots are not.
Website check, a client site · 2 Aug 2026, 00:56 to 01:46 UTC
Second round, 6 Aug 2026: llms.txt added, 89 → 100. Both pairs are on file as dated before/after records.
AI answer check, instlyscout.com · in progress
Baseline: named in 0 of 20 answers. Five answer pages published and submitted to Google and Bing on 30 Sep 2026. The re-check runs no earlier than four weeks later, and the result goes on this page whichever way it lands.
One mention is one twentieth of a check, so a single reading can move five points with nothing changed. We report counts, not percentages, and call a trend only on repeated movement.
We run this on our own sites.
Same two instruments, same dates shown. A site can pass every technical check and still be named nowhere; that gap is the whole reason this platform exists.
| Site | Website check | Named in AI answers | Named instead |
|---|---|---|---|
| instlyscout.com AI visibility software | 100 30 Sep 2026 | 0 of 20 29 Sep 2026 | Semrush 4, Profound 4, Scrunch 4, Otterly 3 |
| ipmicompare.com IPMI comparison | 100 30 Sep 2026 | 3 of 20 28 Sep 2026 · cited as a source | Pacific Prime 4, International Citizens 3, Squaremouth 3 |
| instlytechnologies.com Insurance technology | 73 30 Sep 2026 · no llms.txt, no schema, short description | 0 of 15 8 Sep 2026 · 3 engines, ChatGPT not run | Groundspeed 3, Cytora 3, Vertafore 3 |
| healthinomad.com Nomad health cover | 100 30 Sep 2026 | not yet measured | n/a |
Website check: the thirteen weighted technical checks above, homepage, run by our measurement platform. Named in AI answers: five buyer questions to four engines, 20 answers, counted by the same matching rules every run. Full Scout report on instlyscout.com: AI readiness 90/100, 8 pages, 29 Sep 2026. Numbers on this page are updated when a new check runs, not on a schedule.
Free scan: the thirteen technical checks, instant. The full report adds the AI answer check for your buyers' questions.