Most Singapore businesses pour their energy into earning backlinks from other websites, while the links sitting inside their own site go almost entirely unmanaged. That is a costly blind spot. Every link from one of your own pages to another is a signal you control completely, and it is one of the clearest paths a crawler, human or otherwise, uses to decide what your site is actually about.
Generative AI platforms add a new reason to care. ChatGPT, Gemini, Perplexity and the rest do not read your website the way a person browsing from the homepage does. They rely on automated crawlers that follow links, and if those links do not lead somewhere sensible, large parts of a site can simply go unseen.
Why AI Crawlers Don’t Browse Like Googlebot
Googlebot has spent over two decades patiently working through websites, following sitemaps and revisiting pages on a schedule. AI crawlers behave differently, and the gap is widening fast. Cloudflare’s analysis of crawler traffic across its network found that AI and search bot activity rose 18% on a fixed set of customers between May 2024 and May 2025, with GPTBot’s share of AI crawler requests climbing from 2.2% to 7.7% and ChatGPT-User requests up by more than 2,800% over the same period. That growth is uneven and selective. Some bots surge while others, such as Bytespider, pull back sharply, and Cloudflare found that of the top 10,000 domains with a robots.txt file, only around 14% had rules written specifically for AI bots. A crawler that visits irregularly depends far more on a clear internal link structure to find your important pages than one that re-crawls everything on a fixed schedule.
The Internal Link As A Trust Signal
Internal linking for AI search follows the same underlying logic that has always applied to traditional search: a page receiving more internal links, with more varied and descriptive anchor text, reads as more important than one sitting at the edge of your site with nothing pointing to it.
Cyrus Shepard’s research firm Zyppy analysed 23 million internal links across roughly 520,000 URLs on 1,800 websites and found a striking pattern. Pages with 40 to 44 internal links averaged about four times the organic clicks of pages with fewer than five, before the benefit tapered off beyond around 45 to 50 links. Anchor text variety correlated strongly with higher clicks, and pages carrying at least one exact-match anchor saw at least five times more traffic than pages without one.
The study is correlational, and its own author calls it directionally useful rather than definitive proof. Even so, the pattern fits how retrieval models seem to behave too: context and emphasis, repeated from multiple angles across a site, build a stronger signal than a single mention ever could.
Orphan Pages Are Invisible To Everyone
An orphan page, one with no internal links pointing to it at all, is effectively invisible, however well it answers a question your customers are already asking a chatbot. Service pages added outside the main navigation, old campaign landing pages, FAQ pages bolted on by a plugin, and seasonal content nobody cleaned up afterwards are the usual culprits, and each one represents content an AI platform has no reliable way to discover.
Building Topic Clusters AI Models Can Follow
A deliberate structure solves this far better than linking pages at random. Group related content around a pillar page, link outward to supporting articles, and link back from each supporting article to the pillar and to its closest siblings. This gives both traditional crawlers and AI retrieval systems a map of how your topics relate to one another, which is exactly the kind of context that helps a model treat your site as a credible, well-organised source worth citing.
The anchor text you choose carries as much weight as the link itself. A generic “read more” tells a crawler nothing about the destination page, while a descriptive phrase drawn naturally from the sentence around it does real work. A tightly linked cluster of pages achieves little, though, if your robots.txt rules are blocking the bots that would otherwise follow those links through to the content.
A Practical Internal Linking Audit
Most sites benefit from a structured look at how their pages connect before making changes. A useful audit typically covers:
- Running a full crawl to identify orphan pages with zero internal links pointing to them
- Checking click depth, so no important page sits more than three or four clicks from the homepage
- Reviewing anchor text across the site for generic phrases that could be replaced with descriptive ones
- Mapping pillar and cluster relationships to confirm supporting pages link back to their pillar
- Auditing navigation and footer links for outdated pages that no longer deserve sitewide prominence
- Flagging any new page published without at least one contextual internal link pointing to it
Internal Linking Is One Part Of A Larger System
Fixing even a handful of these issues tends to produce a noticeably clearer site structure within a few weeks, particularly on small and mid-sized sites where a few pages often carry a disproportionate share of the linking weight. None of this replaces the other fundamentals already covered on this blog, either. Entity authority, structured data and clean crawler access all work alongside internal linking, not instead of it, and a site can have beautifully structured links and still struggle if its robots.txt file is turning crawlers away before they reach those links at all.
What makes internal linking worth prioritising now is that it stays entirely within your control. You cannot force another website to link to you, and you cannot dictate how often an AI platform chooses to recrawl your domain, but you can decide, today, which pages on your own site deserve more attention and make the links reflect that.
Getting your site’s internal structure right is foundational work, and it pairs naturally with a wider AI search strategy built around how your content gets found, understood and ultimately cited. If you would like a second pair of eyes on how your pages connect, feel free to get in touch with our team.