Orphan pages and AI crawlers: what to check on your site

An orphan page is a page on your site that no other page links to. It can still exist in your sitemap, it can still load in a browser if you know the address, and it can still be perfectly good content. But a crawler that discovers pages by following links has no path to it, and a crawler with limited time may never go looking.

This guide explains how to find orphans and weak pages, what you should and should not conclude from them, and how to check whether AI crawlers reach the pages you care about.

What is and is not known

Let us be careful here, because this topic attracts overclaiming. It is reasonable to say that pages with few internal links are harder for any crawler to discover and are a poor signal of importance on your own site. It is not established that orphan pages receive fewer AI citations, and you should not tell stakeholders that fixing them will raise citations. What you can say is that you are removing an obstacle to being crawled, and then measuring whether crawl behavior changes.

The distinction matters because it tells you what to measure. Your goal is not "fewer orphans." Your goal is "important pages are reached by the bots that matter, and we can show it."

Four kinds of weak page

When you audit internal linking, sort pages into these groups:

Orphans. Zero internal links point to them.
Weak pages. One or two internal links, often from a single index or archive page.
Redirect and error targets. Internal links that lead to a redirect chain or a broken page, wasting the link.
Sitemap-only pages. Listed in the sitemap but reachable by no navigation path.

Different fixes apply. Orphans that matter need links added from relevant pages. Orphans that do not matter (old campaign pages, thin duplicates) may need to be removed or redirected. Broken targets need their links corrected.

How to find them with no special tool

Export your sitemap URLs. Most CMSs can list them.
Crawl your site from the homepage with any crawler you already use, and export the list of discovered URLs along with their inlink counts.
Compare the two lists. URLs in the sitemap but missing from the crawl are your orphan candidates.
Sort by inlink count. Anything with one or two is a weak page.
Cross-check with Search Console. A page that gets impressions but is barely linked internally is an easy win, because search already sees value in it.

This is a perfectly workable routine. It is also a manual comparison you have to repeat each time the site changes, and a plain crawl cannot tell you whether an AI bot ever visited the pages in question.

Fixing the pages that matter

For each important orphan or weak page:

Add a contextual link from two or three relevant, already-popular pages. Use descriptive anchor text, not "click here."
Add it to a hub or category page if one exists.
Check that it is in your sitemap and returns a normal status code.
Make sure your robots.txt does not block it. Our guide on which bots your site is open to covers that check.

Then wait and measure. Which brings us to the part most audits skip.

Check whether bots actually arrive

Internal links are an input. Crawler visits are the output. If you fix linking but never look at crawl logs, you are guessing. Server or CDN logs show which AI crawlers, such as GPTBot, OAI-SearchBot, ClaudeBot, or PerplexityBot, requested which URLs, and you can compare before and after. Our post on crawl budget for AI crawlers explains why some pages get visited repeatedly while others get ignored.

Where Promptwatch fits

Promptwatch brings the linking audit and the crawler evidence into one place. As of the October 28, 2026 changelog, Site health has a Linking view with these parts:

A map of your site by URL path. Orphan pages, pages with only one or two inlinks, and redirects or errors each get their own color. Click a section to zoom in, or filter to the weak spots.
A Linking tab on every page. It shows the pages linking to it on the left, the pages it links to on the right, and broken links in red.
Page sources. Every page list shows where a page came from: your sitemap, a link from another page, Search Console, or a site map crawl. That is exactly the comparison we did manually above.
Broken internal link issues. Each broken target is listed with its status code and anchor text.
Sitemap-only audits. Turn them on to check only the pages in your sitemaps. New projects start with it on, and pages that drop out of every sitemap leave the audit after the next full sync.

There is also a weekly audit email every Tuesday, covering the health score, issues fixed since the last audit, and the top issues AI crawlers may trip over.

On the creation side, the Content Agent now looks for relevant pages on your site before it starts writing, so internal links are part of the article from the first draft. Per the changelog, pages with many Search Console impressions but few clicks are suggested first. It also stopped linking the homepage every time the brand name comes up.

The crawler side is Agent Analytics: real-time, IP-verified AI crawler logs with a crawl-to-citation path. That is where you confirm the fix worked. Be aware of the plan gating: crawler logs start on Professional ($245/mo), and Essential has none. The linking map and site audit are a separate matter from the logs, so you can start with the audit even if logs are not in your plan yet.

We would still make the same honest caveat as above. Promptwatch shows you the structure and the crawl behavior. It does not promise that connecting an orphan page produces citations. What it gives you is a way to test the idea: link the page, then watch whether bots start requesting it.

For the technical audit as a routine, see our guide to the weekly site audit and its 21 checks and our technical optimization guide for AI crawlers.

A one-hour plan

List your ten most commercially important pages.
Check each one's inlink count. Flag any with two or fewer.
Add two relevant contextual links to each flagged page.
Note the date and the current crawl activity for those URLs, if you have logs.
Check again in three to four weeks.

If bots start visiting, you have evidence that the structure was the obstacle. If they do not, you have ruled out one cause cheaply, and you can look at access rules, page speed, and content quality next.