In June 2025, Google announced the introduction of MUVERA (Multi-Vector Retrieval Architecture), a new component in its search software architecture.
The new algorithm is based on the vector representation of information - essentially the same approach used by LLM models and Generative AI - to connect concepts and data. This effectively integrates and gives even greater weight to the Google Knowledge Graph, which was first introduced in 2012, in the process of generating search results for every query.
Google's new approach is therefore a further commitment to the semantic representation of information, moving beyond simple keyword matching to focus on expressing meaning and intent. This also presents a new challenge for online content publishers, as their content is no longer treated as mere text but as semantic entities within the knowledge graph.
This highlights the need to anchor own contents within Google's semantic database. This means not only relying on Google to correctly interpret each page but also actively using the structured data offered by the Schema.org ontology (officially supported by Google) to explicitly define desired concepts and express relationships with Wikidata entities, as Wikidata is the world's largest public knowledge graph, and Google's Knowledge Graph draws from it to establish common reference points.
It is for this reason that Semantic Builders, a pioneer in semantic SEO, offers its "Named Entity Recognition" service. The tool simplifies and speeds up the process of labeling published online content, giving it an explicit semantic context that crawlers (from Google, as well as AI chatbot providers like OpenAI or Perplexity) can assimilate and use to place each page within their respective graphs. Ultimately, this enriches the vocabulary they use to provide coherent answers to user requests.
The tool is provided as an API and requires integration with your web platform, but we are always happy to offer practical and operational support to clients who want to explore the opportunities offered by this new model.
Also in the field of Semantic SEO, we would like to suggest the service offered by our partner, Trending On Wiki. Available in Spanish and in Italian, their solution adds an extra layer of semantic contextualization. It further simplifies the identification of relationships with Wikidata entities, helping to reduce "noise" and allowing you to focus only on the concepts that are important to your editorial strategy.
