INCREV Query Match Tool - Applying Sentence-BERT Algorithm to analyze Topical Authority for onpage content and link relevance. DOI: 10.5281/zenodo.17571849 (David Vesterlund, IncRev.co)

IncRev SEO Research team has developed QueryMatch, a tool that uses the SBERT Algorithm to measure similarity score between keywords and content. This has several applications within SEO Analysis and AI Search Analysis like: Topic clustering for Topical Authority, opitimize Topicality in articles for onpage SEO, Measure relevance for potential link building oppertunities, audit existing backlink relevance and many more.
In this article we study how Sentence-BERT (SBERT) sentence embeddings and cosine similarity can be operationalized for SEO tasks.
We present Query Match, an INCREV system and tool that scrapes target pages, 
vectorizes content at paragraph granularity, and automatically rewrites low-similarity sections to surpass configurable thresholds.
We connect our approach to established distributional semantics (PMI factorization) and recent transformer-based encoders, and we report design heuristics for reliable deployment in multilingual settings. 
We also show the high correlation (88-97% depenging on model) in vector spaces between SBERT and both Googles and Open AIs embedding algorithms as well as the linearity of correlation across the vector space. This high correlation indicates SBERT and INCREVs QueryMatch to be a good proxy for understanding hos search engines and AI engines understand and measure topicality and relevance between search queries and the competing subpages in the SERP.