What do agent-readiness scans measure? Reading a score in context
When a site receives a high technical score, it is tempting to expect frequent citations in AI answers too. The two results answer different questions. One reports the signals a particular scanner checks; the other requires observing which sources actually appear for particular user prompts.
Agent-readiness scans look for features that may help automated tools discover a site or access its content. The checks and the meaning of the score depend on the tool and profile selected.
Which signals are checked?
Is It Agent Ready says it checks discovery signals such as robots.txt and sitemaps, alongside Markdown delivery, bot access rules and various agent protocols. It offers different profiles, including one for content sites. A blog's lack of a commerce protocol does not by itself mean its articles cannot be read; the relevance of a check depends on the site's purpose.
Finding a robots.txt file does not establish that every URL is accessible to every bot. Server responses, redirects, authentication and security layers can affect the outcome. The details next to a score often tell you more than the score alone.
What does the RadioTEDU comparison show?
In research conducted on 8 November 2026, the Is It Agent Ready content profile returned level 5/5 for radiotedu.com, level 2/5 for technopat.net and level 1/5 for donanimhaber.com. These are technical profile results from that date, not comparisons of organic traffic, search rankings or AI citations. Results can change when scanner rules, server responses or redirects change.
A different number answers a different question. Semrush’s Technopat domain overview showed an estimated 3.25 million monthly organic-search visits for September 2026. That is a third-party estimate for the domain, not the publisher’s analytics, the traffic of an individual article or a count of AI citations. The technical scan was conducted on 8 November, while the estimate concerns September. Even read together, the two figures do not establish a causal relationship or the site’s GEO performance.
One AI answer illustrates the limits of the score. In a single Perplexity test on 8 November 2026 asking about contemporary art and culture publications in Türkiye, ArtDog Istanbul and e-skop appeared among the visible sources. Their homepages both received 0/5 in Is It Agent Ready’s content profile on the same day. This is an observation from one prompt, one platform and one session; it does not establish citation frequency, performance across other prompts or why those sources were selected. Public access to the session was not verified.
Crawler names serve different purposes
Listen while exploring this story
Tune into campus radio with non-stop independent music and culture.
A scan may list several bot names, but they do not all have the same job. According to Google’s crawler documentation, Google-Extended controls certain Gemini training and grounding uses in Gemini Apps and Vertex AI. It does not determine inclusion in Google Search or act as a Search ranking signal. It should not be confused with Googlebot’s Search crawling. Google-Extended does not have a separate HTTP user-agent string; Google uses it as a robots.txt control for content fetched by existing crawlers.
OpenAI’s crawler documentation associates OAI-SearchBot with ChatGPT’s search features and GPTBot with the potential use of crawled content to train foundation models. It describes these as independent settings: allowing search crawling does not require allowing training crawling. ChatGPT-User serves some user-initiated actions and is not the automatic search crawler.
OpenAI says a site that opts out of OAI-SearchBot will not be shown in ChatGPT search answers, though it may still appear as a navigational link. Robots.txt rules may not apply to user-initiated actions made with ChatGPT-User. Those distinctions matter when deciding which type of access to allow.
For search access, OpenAI recommends checking both robots.txt and whether the host or CDN allows its published IP ranges. These settings do not guarantee ranking or citation, and this example does not verify RadioTEDU’s current firewall or CDN configuration. OpenAI notes that its systems may take about 24 hours to adjust to a robots.txt change for search results.
An inaccessible scan is not a zero
When a domain redirects, record the final target that the scan actually reached. If an HTTP error or access restriction prevents a result, mark the test failed or unmeasured rather than assigning zero. The date, profile, target and failed checks give a comparison its proper context.
Google's robots.txt introduction explains that robots rules manage crawling and do not by themselves guarantee that a URL will never appear in search results. A scanner's robots check cannot establish the actual index status of an individual article.
What belongs beside the score?
Checking real pages for access and index status, then sampling dated Turkish and English prompts, provides different evidence. Search Console impressions and clicks can also be reviewed separately when access is available. Together these observations offer a richer picture; none alone measures all of “GEO success.”
A high technical result may help identify where to investigate. Its value comes from reading it within its scope and then asking what readers actually find.
