Internal linking is the practice of strategically linking from one page on your site to another. It does three jobs:
- Authority distribution — pages that are linked to from many other pages tend to inherit ranking signal
- Topical clarity — Google reads anchor text and link patterns as signals of what a page is about
- User flow — readers (and crawlers) discover related content
Vector-based suggestion
Most internal-link tools work on string matching: "this page mentions Topic X, link it to your other page about Topic X." That misses pages that are semantically related but use different words.
BCited uses embeddings. We:
- Build a vector representation of every URL on your site (from its path + the queries it ranks for)
- For any cluster, find URLs whose vector is closest to the cluster's centroid
- Surface them as link candidates
The result is suggestions that include pages with no exact word overlap but obvious topical kinship — the connections an experienced editor would make but a topic model wouldn't.
When generating briefs
New briefs automatically include the top 5 internal-link candidates, ranked by semantic similarity weighted by target impressions. The result is a new page that lands in the right neighborhood of your site from day one.