Has anyone actually seen stable gains from AI-assisted internal linking at scale?

[mercer]

I’ve been testing a few AI-assisted internal linking workflows across two niche sites over the last couple of months, and I’m still not fully convinced the results are as clean as some people claim.

The basic setup was:

– export all existing URLs and target keywords
– use an AI workflow to identify contextual link opportunities
– manually review the suggested anchors
– push links into older posts with decent topical overlap
– track changes in impressions, average position, and crawl activity

On paper, it sounds like one of those easy wins. In practice, I’m seeing a mix of outcomes.

What’s working:
– pages with weak internal support do seem to pick up impressions after being linked from stronger topical clusters
– orphaned or near-orphaned pages benefit the most
– AI is decent at spotting semantically related anchors that a quick manual pass might miss
– on larger sites, it saves a lot of time compared to doing everything by hand

What’s less convincing:
– some suggested links are technically relevant but feel forced in the actual paragraph context
– a few pages got more internal links but no meaningful ranking movement
– in one case, the AI kept over-linking to the same “commercial” pages instead of balancing informational pages
– I’m not sure how much of the improvement is from internal links themselves versus just refreshing older content at the same time

My current view is that AI internal linking is useful as a scaling layer, but not something I’d trust without manual review. The quality threshold still matters a lot. If the anchor feels unnatural, I think it can dilute the page rather than help it.

I’m curious how others are handling this:

1. Are you using AI to suggest internal links, or only to automate insertion?
2. Have you seen measurable ranking gains, or mostly better crawl/discovery?
3. Do you limit links by topic cluster, or let the system find any semantically related page?
4. For affiliate sites, are you prioritizing money pages or building more informational support first?

Would be good to compare notes with people running this on real sites, not just demos. If there’s a repeatable workflow here, I think it’s probably in the combination of AI suggestions + strict manual filtering + a sensible site architecture, not in automation alone.

[Author]

Posts