I’ve been testing a few different internal linking workflows across niche sites over the last couple of months, and I’m starting to think the “AI will solve internal linking at scale” pitch is a bit overstated.

What I’ve tried so far:

– exporting existing URLs and using an LLM to suggest contextual link placements
– running link suggestions based on target keyword clusters
– using automation to insert links into draft content before publishing
– manually reviewing everything before it goes live

The results have been mixed.

On the positive side, AI is decent at identifying obvious opportunities:
– supporting articles that should point to a money page
– related informational posts in the same cluster
– orphaned pages that need some internal visibility

But where it gets shaky is:
– over-linking the same pages repeatedly
– suggesting links that are technically relevant but awkward in context
– missing intent differences between pages that look similar on paper
– creating patterns that feel unnatural if you scale it too aggressively

I’ve also noticed that internal linking only seems to move the needle when the site already has:
– decent topical structure
– enough content depth
– clean navigation and category architecture
– pages that are already close to ranking

If the site is thin or poorly organized, AI-generated internal links don’t seem to do much beyond making the structure look more “optimized” in spreadsheets.

My current view is that AI is useful as a helper, not a replacement. The best setup I’ve found is:
1. build the topical map manually
2. let AI suggest link opportunities
3. review for intent and readability
4. automate only the low-risk inserts
5. keep an eye on over-optimization

I’m curious how others are handling this.

Has anyone here found a workflow that actually scales internal linking without turning the site into a mess?
And more importantly, have you seen measurable ranking improvements from AI-assisted linking, or is it mostly just a time-saver?

Would be interested in examples, especially from people running affiliate sites or larger content libraries.