Skilljar by Gainsight: Make Your Academy Discoverable by AI Assistants
Your customers are increasingly turning to AI assistants for answers; how to configure a feature, which course to take, where to start. When your Academy content isn't readable by those tools, the answer comes from somewhere else: a competitor's page, an old forum thread, or a best guess.
How AI tools discover and cite content is still an evolving space, but we didn't want that to stop us from giving your Academy a seat at the table.
What AI Discoverability Does
Our first release of Academy AI Discoverability makes your catalog readable by AI assistants and gives you one setting to control whether AI crawlers are allowed to index your content.
When enabled, AI assistants crawl structured metadata on course pages. Your courses carry machine-readable data that tells AI tools exactly what they're looking at: a course, who offers it, what it teaches, and whether enrollment is required. Instead of treating your catalog as just another web page with words on it, AI assistants can understand and accurately represent your training.
How to Turn It On
In your Dashboard, open Domain Settings for each domain you want to enable. Find the AI Discoverability section and check "Allow AI crawlers to index this academy."
What Happens Behind the Scenes
The checkbox controls three files:
| File | What it Does | Can You Change It? |
| robots.txt | The file that tells crawlers where they may and may not go. Unchecked (AI Discoverability off): Skilljar adds a Disallow: / block for 14 known AI crawlers to the top of the file. One block per crawler, each blocking that bot from the your Academy: User-agent: GPTBot Disallow: / User-agent: ClaudeBot Disallow: / ... (12 more AI crawlers) ...
Checked (AI Discoverability on): those 14 crawler-specific blocks are removed, and the file reverts to just the baseline shown above. In other words, turning the feature on doesn't add anything to the file, it removes the AI-specific blocking rules that are there by default. | Only via the checkbox |
| llms.txt | This plain-text file is a machine-readable index of every public course and learning path, including titles, descriptions, and links. It's generated from your live catalog and updates whenever you publish changes. | No |
| sitemap.xml | Standard list of pages for search engines. Note: While AI Discoverability is turned off, /sitemap.xml returns a 404 error for all requests, including Google searches. If you rely on your sitemap for SEO, enable AI Discoverability for the relevant domains. | No |
What llms.txt Exposes to AI Assistants
llms.txt will never expose lesson content, video transcripts, quiz questions, or SCORM files. AI tools learn what training you offer, not what's inside it. Private academies return a 404 instead of a listing.
llms.txt publishes your Academy name and the title, description, and link for each public course and learning path.
A course appears in /llms.txt when it's published, visible on your catalog, and your Academy is checked to allow AI Discoverability. The description shown in AI results is pulled from the course's short description field.
A Quick Note on Best Practices
If you’re looking to improve how AI tools describe your courses, we recommend keeping the short description field up-to-date and setting a cadence for regular updates. This is a feature worth planning for, not just switching on.
For years, discoverability meant optimizing for a ranked list of blue links, getting the keywords right, and hope for page one. AI answer engines work differently: there's no page two. An assistant either has enough signal to cite you, or it reaches for whatever else it can find. This raises the bar on freshness and clarity, and lowers the bar on volume. A handful of clear, current course descriptions will outperform a large but stale catalog.
It also changes what "traffic" means. A citation in an AI assistant's answer may never show up as a click in your analytics, but it's still a customer forming an impression of your training before they land on your site. Treating short descriptions as living copy, reviewed on a schedule, is becoming as routine as maintaining meta descriptions was for search.
If you want a structured way to approach AEO, check out our 90-Day AEO Checklist, a step-by-step guide to thinking about your strategy before rolling this out more broadly.
One area still ahead of us: an authorized learner's own AI assistant pulling course content directly on the learner's behalf. We're actively thinking through that mechanism.
What to Expect After You Turn It On
Turning this on doesn't guarantee your Academy shows up in ChatGPT tomorrow; how AI tools decide what to crawl and cite is still evolving. But here's what we know:
- Crawling takes days to weeks. Every AI provider works on its own schedule, and there's no way to speed it up.
- Citation isn't guaranteed. You're making your content readable, not buying placement. Whether an AI tool cites you is ultimately decided on the provider's crawl behavior.
- Recency matters. AI tools tend to favor recently updated content, so an actively maintained catalog is more likely to be picked up.
We'll continue to share what we learn about crawler behavior and citation patterns as this space develops. Visit our Help Center article for additional setup instructions and FAQs.