Guide · Online courses

How do online courses get recommended by AI?

Updated 2026-11-09 · kagentx

Short answer

First make the course page readable to crawlers, then write the course information as body text. On 2026-11-09 we read the robots.txt of five course platforms: Coursera, Teachable and Podia name GPTBot, ClaudeBot, PerplexityBot and others, and course pages are Allow for them; Skillshare names none; Udemy's robots.txt returned a 403 verification page and is unverified. In the two course pages we could open, the syllabus appears as text in the raw HTML. A course page should state who it is for, prerequisites, syllabus, total length, price and instructor. Google has removed its Course info documentation and Course list remains; whether AI assistants read Course markup is not stated by the vendors and is unverified.

Do course platforms block AI crawlers in robots.txt?

Of the five we checked, three name AI crawlers and allow them, one names none, and one has a robots.txt we could not read. On 2026-11-09 we fetched each robots.txt with curl and judged by course-page paths; this is a reading of the robots.txt text, not the behaviour of real crawlers, and a CDN or firewall may block separately; see How do you write robots.txt for AI crawlers?.

PlatformGPTBotClaudeBotPerplexityBotNotable in robots.txt
Coursera (www.coursera.org)Allow(点名)Allow(点名)Allow(点名)These three crawlers are additionally blocked from /lecture/, /search and more; OAI-SearchBot and ChatGPT-User have their own group that does not block /lecture/; both groups allow /llms.txt
Teachable (www.teachable.com)Allow(点名)Allow(点名)Allow(点名)Names ten AI-related crawlers with the same rules as *: only back-office, login and search paths are blocked
Teachable 学校子域 (anna-geiger.teachable.com)AllowAllowAllowThe * rules block only back-office paths such as /secure/, /admin/ and /manage/
Podia (www.podia.com)Allow(点名)Allow(点名)Allow(点名)Also names OAI-SearchBot and Google-Extended, all Allow: /; podia.com 301-redirects to www first
Skillshare (www.skillshare.com)AllowAllowAllowNames no AI crawler; the * rules block search, account and checkout paths
Udemy (www.udemy.com)Not verifiedNot verifiedNot verifiedrobots.txt returns a 403 verification page; the content cannot be read

"Allow (named)" means robots.txt has rules that name this crawler; "Allow" means it is not named and the * rules apply. Udemy's robots.txt returned a verification page to our curl; that does not mean it blocks AI crawlers, and we did not try to get around it, so we mark it unverified.

Is the syllabus in the course page's raw HTML?

In the two examples we could open, yes. On 2026-11-09 we fetched the raw HTML with curl (no script execution) and looked at the syllabus and structured data:

Course pagePage without logging inSyllabus in the raw HTMLStructured data
Coursera /learn/machine-learning200Yes: the titles and descriptions of Weeks 1–3 are in the body textHas Course (with 3 syllabusSections, educationalLevel, coursePrerequisites, and hasCourseInstance with courseWorkload and instructor, among others), plus FAQPage and BreadcrumbList
Teachable 学校课程页 (anna-geiger.teachable.com/p/reach-all-readers)200Yes: each module's lesson titles and durations (such as 7:24) are in the body textNo JSON-LD and no meta description; it has a canonical
Skillshare 课程页403Unverified: curl returned a verification pageNot verified
Udemy 课程页Not verifiedUnverified: robots.txt itself was blocked by a verification page, and we did not try the course pageNot verified
Podia 课程页Not verifiedUnverified: we found no publicly openable exampleNot verified

How to read it: all we can confirm is that the syllabus appears as text in the raw HTML. Whether AI assistants read that text or Course markup is not stated by the vendors and is unverified. For the effect of script rendering on crawlers see Do AI crawlers render JavaScript?.

Commands to check your own course page:

# Is the syllabus in the raw HTML? Replace the quoted text with a lesson title from your page.
curl -s https://example.com/courses/intro/ | grep -c "Lesson 1: Light"
# Is there Course markup?
curl -s https://example.com/courses/intro/ | grep -o '"@type": *"Course"' | head -1
# What does robots.txt say about the AI crawlers?
curl -s https://example.com/robots.txt | grep -i -A3 -E "gptbot|claudebot|perplexitybot|user-agent: \*"

What text should a course page contain?

These are what buyers ask when comparing courses, so write each as body text, not only in a video or image. This table is our own way of organising the questions buyers ask, not a requirement from any vendor:

What to writeHow to write itExample
Course name and one-line outcomeMake the first sentence say "what you can do after finishing""After this you can shoot product photos good enough to list"
Who it is forState the level and the role"Complete-beginner online shop owners"
PrerequisitesWrite the equipment, software and prior knowledge needed; write "none" if there is none"A phone; no editing software needed"
SyllabusWrite a title for every chapter and lesson, and the length if you have it"Lesson 2: Framing, 12 minutes"
Total length and formatTotal length, self-paced or live, and whether there is a deadline"3 hours in total, self-paced"
PricingWrite the number, the currency and the update date; see How should a pricing page be written so AI answers your price correctly?"$49, updated 2026-11"
InstructorName plus a sentence or two of real experience, linked to a profile page"Jane Doe, ten years of product photography"
VerdictSay who, when, and link the original words; see How should testimonials and case studies be written for AI to cite them?"2026-10, link to the original post"

Each specific question ("can a beginner follow it?", "do I need equipment?") can get its own section, written as described in Is FAQ structured data (FAQPage) still useful?; for pages organised by scenario see How to write "X for Y" use-case pages that AI will cite.

How do you write Course structured data?

Google's current position first. We viewed Google's Course list (Course) structured data documentation on 2026-11-09: to be eligible for the course list rich result you mark up at least three courses; name and description are required (description has a display limit of 60 characters) and provider is recommended; a list page uses ItemList, and each item needs position and url; each course can have its own detail page, or all can sit on one page. The feature is offered in English in all regions where Google Search is available, and Google does not guarantee it will be shown.

Another thing: Google has removed its Course info documentation. Its update log says that on 2025-09-09 it removed the documentation for course info and a few other structured data types because they are no longer shown in Google Search results; a 2025-06 blog post also said these types would be phased out. So do not follow old tutorials about Course info, and do not expect extra display from it.

The schema.org/Course vocabulary itself is still there, and Coursera's course pages still use it (including syllabusSections and hasCourseInstance). Below is a single-course example written with schema.org properties, meant to put a machine-readable copy of facts already in the body text:

<script type="application/ld+json">
{
  "@context": "https://schema.org",
  "@type": "Course",
  "name": "Intro to Product Photography",
  "description": "Six lessons on lighting and framing for small online shops; about 3 hours in total.",
  "provider": { "@type": "Organization", "name": "Example Studio", "sameAs": "https://example.com/" },
  "educationalLevel": "Beginner",
  "coursePrerequisites": "A phone or camera; no editing software needed.",
  "syllabusSections": [
    { "@type": "Syllabus", "name": "Lesson 1: Light", "description": "Window light and one lamp." },
    { "@type": "Syllabus", "name": "Lesson 2: Framing", "description": "Angles that work for small products." }
  ],
  "hasCourseInstance": {
    "@type": "CourseInstance",
    "courseMode": "Online",
    "courseWorkload": "PT3H",
    "instructor": { "@type": "Person", "name": "Jane Doe" }
  }
}
</script>

Three rules: every number and text in the markup must match what the page shows; do not add ratings that are not on the page; whether AI assistants read it is not stated by the vendors and is unverified. For the general judgement see Does structured data help AI search?.

Platform page or your own site: which should be the main page?

Our advice is to make the course page on your own site the main page and treat the platform page as a distribution channel. The reasons come from what we observed, not from any vendor statement:

  1. 1. You can change little on a platform page

    The Teachable school course page we looked at has no meta description and no structured data; Skillshare and Udemy returned verification pages to curl. On your own course page, all of these are in your hands.

  2. 2. Write everything on your site's page and repeat the same facts on the platform page

    Take the syllabus, length, prerequisites, price and update date from your site, and copy the same text into the platform description. Use one pricing wording only.

  3. 3. Link from the platform page back to your site and keep the two consistent

    Put your site's course page URL in the platform description. For how to treat near-identical pages, see Canonical tags and duplicate content: how to set them up on a multilingual site so pages are not treated as duplicates.

  4. 4. Test for yourself which gets cited more, the platform page or your site

    Ask AI the buyers' questions (with web search on) and see whether the sources show the platform page or your own site. The vendors do not say which they prefer and we have no evidence, so this is unverified.

FAQ

Do course platforms block AI crawlers?

On 2026-11-09 we read the robots.txt of five platforms: Coursera, Teachable and Podia all name GPTBot, ClaudeBot, PerplexityBot and others, and course pages are Allow for them; Coursera also blocks /lecture/ for GPTBot, ClaudeBot and PerplexityBot. Skillshare names none and follows the default rules. Udemy's robots.txt returned a 403 verification page and is unverified.

Does every course page need Course structured data?

Not necessarily. Nobody official says AI assistants read it, and we have not verified that. Google's Course list documentation requires marking up at least three courses, to compete for the course list rich result. The value of a single course page is mainly a machine-readable copy of facts such as syllabus, length and prerequisites, kept consistent with the body text.

What happened to Google's Course info documentation?

Google removed it. Its update log says that on 2025-09-09 it removed the documentation for course info and a few other structured data types because they are no longer shown in Google Search results. The Course list documentation remains. The schema.org/Course vocabulary itself is still there, and Coursera's course pages still use it.

If my course is on a platform, do I still need my own site?

We recommend it. You may not be able to change robots.txt, meta or structured data on a platform page: the Teachable school course page we looked at has no meta description and no JSON-LD. On your own course page you can write the full syllabus, price and prerequisites. Which one AI cites more often is not stated by the vendors and is unverified.

Is it OK to put the syllabus only in a video?

Not recommended. On the Coursera and Teachable pages we looked at, the syllabus appears as text in the raw HTML. Text inside videos and images cannot be read by crawlers, so also write it as body text.

Sources

kagentx

Can crawlers read your pages?
Let the evaluation tell you

The evaluation checks whether crawlers can get in and whether the text is served directly, then asks ChatGPT, Perplexity and Claude (with web search on) your buyers' questions and sees whether you are mentioned. The Basic plan is $1, and results usually take 2–3 minutes.