# How to Find Orphan Pages on a Business Website By AI Visibility editorial team AI-assisted article. A business has a useful warranty page. Staff send it to customers by email, and it appears in the XML sitemap, but a customer starting on the website cannot find it through the service pages or navigation. The address works when someone already knows it. The missing piece is an ordinary path from relevant content. Start with an inventory of pages the business knows about, then compare it with the pages reachable through internal links. Investigate the difference before assigning a label or adding links. The AI Visibility internal linking guide covers how to design helpful connections. This walkthrough shows how to find valuable pages that may be missing from those connections and check whether a new route works. AI Visibility internal linking guide (/article/internal-links-for-ai-search/) Define what counts as reachable Choose the site boundary and one or more public starting pages. For a small business, the homepage is a practical baseline. A site with different hosts or language sections may need more than one starting point. Record the version you tested, such as the public English site in a clean session. Use a definition that someone else can reproduce. Distinguish a page with no incoming link from a page that is linked inside a disconnected group. Two old campaign pages might link to each other while no path from the public site reaches either one. A simple incoming-link count can miss that case. Check both whether another page links to the target and whether the link can be reached from a useful starting point. Ignore links a page makes to itself when measuring whether another source leads there. Finding What it means in the audit What to check next In the sitemap and reachable from home Listed and linked in the tested graph Whether the route makes sense for customers Listed but no other page links to it Candidate orphan within the collected scope Inventory completeness and intended audience Linked from a disconnected page group Has inbound links but no path from the start Which relevant hub should connect the group Direct link works after a campaign Page is available to someone with the address Whether its purpose calls for permanent navigation URL redirects or is unavailable Inventory entry needs separate review Preferred address or intended retirement Google explains that pages can usually be found when they are reachable through menus or links on other pages. A sitemap provides another discovery route for search engines. Keep sitemap membership and internal reachability as separate observations. A public page reached by a social profile, a QR code or a saved email may still have no route from the business’s public navigation. Google sitemap overview (https://developers.google.com/search/docs/crawling-indexing/sitemaps/overview) Compare independent inventories with the link graph Start with an export from the publishing system if you have one. Add the sitemap, known landing pages and approved reports that contain URLs people actually used. Keep the source of each entry. An analytics report can reveal a campaign URL, while a sitemap may list only pages meant for search discovery. Check status and final address so redirects and unavailable URLs do not inflate the group of valuable pages that appear unlinked. Use the canonical or preferred URL only after checking whether it represents the same content. Extract ordinary anchors with href values from the source pages you can legitimately inspect. Resolve relative links against the document base and record source and destination. Keep query parameters when they distinguish content, and remove fragments for the document-level graph while retaining them for a separate link check. Google’s crawlable-link documentation gives an anchor with an href as its reliable starting point. If the site adds links in a browser or hides important navigation in a menu, check representative rendered states separately. Google crawlable link guidance (https://developers.google.com/search/docs/crawling-indexing/links-crawlable) To compute reachability, start with the homepage and mark each eligible destination it links to. Then follow the outgoing links from those pages, adding each new destination only once. The resulting count tells you how many link steps were needed in the captured graph. Compare the reached set with the inventory. Check the actual reason for each difference before calling a page orphaned. For example, a branch could be missing from the crawl because it is on another host or because the audit stopped before following archive navigation. Consider an illustrative inventory with a homepage, a service hub and three guide pages. The homepage links to the hub and the hub links to the first guide. The other two guides link to each other. A reachability check starting at home finds the hub and the first guide. A count of incoming links alone can make the two disconnected guides look healthy. Adding a useful link from the hub to the appropriate guide connects the group while giving the customer a reason to follow it. For a saved graph, this small Python example shows how to find routes by following only the captured edges. Use the actual homepage and a dictionary whose keys are source URLs and whose values are lists of destination URLs. Keep the eligible inventory separate so the final set difference can be checked against the publishing export. from collections import deque def reachable(start, links, inventory): depth = {start: 0} queue = deque([start]) while queue: source = queue.popleft() for target in links.get(source, []): if target in inventory and target not in depth: depth[target] = depth[source] + 1 queue.append(target) return depth A larger crawl can include public pages found through links even if the sitemap omits them. Record which sources were added and whether following them changed the reachability result. The inventory comparison is only as broad as the sources and edges you collected. See what the public AI Visibility sample showed On 6 November 2026, a public snapshot of AI Visibility’s XML sitemap supplied 42 unique URLs. The audit requested all 42, captured their final response details and collected same-host HTTPS links from the HTML. It counted a document as reached when the homepage could lead to it through those captured links. All 42 returned HTTP 200 directly as HTML. One was the starting homepage, 29 could be reached in one step and the remaining 12 in two steps. No listed page was unreachable in that captured graph. AI Visibility XML sitemap (https://aivisibility.ae/sitemap.xml) · AI Visibility homepage (https://aivisibility.ae/) A second check followed one of the two-step routes in a clean Chrome session at 390 pixels wide. It started on the homepage, followed the link to the second archive page and then followed the link to the service-page writing guide. The browser reached the intended article. That run checked the usability of one actual route, while the graph checked captured links for the sitemap inventory. AI Visibility second archive page (https://aivisibility.ae/?page=2) · Service-page writing guide (/article/how-to-write-service-pages-that-answer-real-customer-questions/) The result is bounded by the inventory. It cannot tell you whether a public URL absent from the sitemap and outside the captured graph exists elsewhere in the publishing system. That is why a first-party publishing export, a permitted analytics report or a list of older campaign URLs can be valuable. Start with a known inventory, document what it covers and put additional sources in a separate comparison when they become available. To repeat the method on an owned site, export the URLs and save a source-to-target list from the public HTML. Treat the start page as depth zero. Review the pages the graph did not reach and the pages that have links only from a disconnected group. Keep the output small enough to review by grouping pages with the same template or source. Repeat a sample in a browser to establish that the links work in the reader’s intended state. Decide which pages need a better route For each candidate, ask who needs the page and where they would naturally look for it. A warranty guide might belong beside repair terms or aftercare. A branch page can be linked from the location directory and from services that operate there. The point of adding a link is to give the reader a useful choice at the moment it matters. The W3C guidance on multiple ways to locate pages gives useful context for considering alternate routes where the criterion applies. W3C guidance on multiple ways to locate pages (https://www.w3.org/WAI/WCAG22/Understanding/multiple-ways.html) Some pages may intentionally stay outside general navigation. An old campaign offer may have expired. A utility screen can be part of a sequence rather than a destination people need to browse to. A duplicate page might need a consolidation decision. Confirm the owner’s intent before adding links. If the content is private, use the appropriate access controls. Internal reachability is a test of discovery paths, and access control is a separate decision. A social channel can give people another way to reach a public page. Planoly’s article about profile links embeds an official post from Instagram’s Creators account. This illustrates a path that starts outside the business website. If the page should also serve people who start from the homepage or a service guide, verify that internal route separately. A campaign link can reach the right page while the regular site journey still lacks a useful connection. Planoly coverage with the official Instagram Creators post (https://www.planoly.com/blog/links-for-instagram-bio) View the selected official Instagram post (https://www.instagram.com/p/CrMAyp4JzSP/) Official Instagram Creators post embedded in coverage of profile links. Add the link, then verify the journey Use an Orphan page review record with the URL, inventory source, HTTP result, observed incoming links, starting page, route length, intended audience and decision. If the page is valuable and a missing route is confirmed, choose a related source with enough context for the reader. Use link text that makes the destination clear. Give the change to the team that owns the source page or navigation component. Retest both the link and the target after publication. Start at the agreed public entry, follow the link and confirm the page displays the expected information. Repeat the inventory comparison and record whether the page is now reachable from the original start. If the business uses a separate mobile menu or a translated route, test the version you intend to support. A direct GET confirms delivery of the destination. The navigation check confirms that a reader can reach it from the intended place. When permitted server logs or search inspection are available, use them to answer additional questions about observed visits. A recorded request for the target cannot tell you whether a visitor came through the new internal route unless the evidence supports that connection. The worked sample used public responses and a browser journey. It had no private logs or platform reports. Google Search Console URL Inspection (https://support.google.com/webmasters/answer/9012289?hl=en) Record any remaining gap in plain terms. For example, a page might now be linked from the service hub while the Arabic navigation is still awaiting a label. This gives the owner a concrete next check and keeps the audit tied to real customer routes. AI Visibility is published by Lunasol, which offers GEO and AEO services. The worked example uses public sitemap, HTML and browser evidence. About AI Visibility (https://aivisibility.ae/about/)