Semantic search, entities and E-E-A-T: how Google and AI engines understand your site

For twenty years, SEO was explained as a matter of keywords: find the phrase, repeat it in the right places, wait. That model is no longer enough. Google and AI engines such as ChatGPT, Copilot or Perplexity do not just match strings of text. They try to work out who is speaking, about what, and whether that source deserves to be trusted.

For a Belgian SME, this is good news. You do not need the biggest budget, you need the clearest identity and the most coherent coverage of your subject. This article defines the key terms (semantic search, entity, Knowledge Graph, topical authority, E-E-A-T), explains how they connect to GEO, and gives you a method you can apply in a few weeks.

An article from our SEO cluster. Links and third-party mentions, the other half of authority, are covered in our article on link building in 2026 for SEO and GEO.

In short

• Semantic search means engines try to understand the meaning of a query and of a page, not only to match words.
• An entity is a distinct, identifiable thing (a company, a person, a product). Engines build their understanding of your site around the entities it describes.
• Topical authority comes from covering one subject in depth and with coherence, not from publishing on everything.
• E-E-A-T is not a ranking factor by itself, but it describes what Google wants to see: real experience, expertise, authority and, above all, trust.
• For GEO, the same foundations apply: a clear identity, identified authors, structured data and corroboration by third-party sources.

What is semantic search, and how is it different from keyword search?

Semantic search is a way of searching that tries to understand the meaning and the intent behind a query, not just the words typed. Google's documentation lists several AI systems that work this way: neural matching connects the concepts in a query to the concepts in pages, BERT helps understand how combinations of words convey meaning, and RankBrain relates words to concepts so that a page can rank even without the exact query terms.

In practice, a page about "Odoo integrator" can answer a query about "ERP partner in Belgium" even if the exact phrase is missing. The reverse is also true: a page that repeats a keyword fifteen times without ever answering the question gains nothing. Write for the question, cover the related concepts and name things precisely.

What is an entity, and what is the Knowledge Graph?

An entity is a distinct, identifiable thing: a person, a company, a place, a product, a concept. Nexuro Digital is an entity. So are Timothy Jacqmin, Odoo, Peppol and Brussels. What matters is that each one is distinguishable from lookalikes and connected to the others: Nexuro Digital was founded by Timothy and Bryan Jacqmin and is an Official Odoo Partner.

The Knowledge Graph is Google's database of such entities and of the relationships between them. Google introduced it on 16 May 2012 with the motto "things, not strings", at the time with roughly 500 million objects and 3.5 billion facts. AI engines rely on comparable representations, built from the web and from the sources they judge reliable.

What is topical authority?

Topical authority is the degree to which a site is seen as a complete and reliable reference on a subject. It is an industry term, not an official Google metric: Google publishes no "topical authority score". But the idea matches how its systems work. Link analysis helps them understand what pages are about, and the helpful content system favours original, people-first content.

In practice, authority comes from depth and coherence: a pillar page, supporting articles that each answer one precise question, and internal links that tie everything together.

What is E-E-A-T SEO, and what does Google say about it?

E-E-A-T stands for Experience, Expertise, Authoritativeness and Trust. The acronym comes from Google's Search Quality Rater Guidelines, the manual used by human raters to evaluate the quality of search results. Experience was added to the former E-A-T on 15 December 2022, and the latest edition of the guidelines we consulted is dated 11 September 2025. It states that trust is the most important member of the family: an untrustworthy page has low E-E-A-T however experienced, expert or authoritative it may seem.

Two clarifications avoid most misunderstandings. First, E-E-A-T is not a specific ranking factor: Google says it uses a mix of signals that help identify content with good E-E-A-T, and raters do not control rankings. Second, the stakes rise on topics that affect health, finances or safety. For an SME, E-E-A-T SEO comes down to making visible who writes, with what experience and on what proof.

What really drives Google ranking in 2026?

Google describes its ranking as a set of automated systems that look at many factors and signals across hundreds of billions of pages. There is no single lever. The systems it documents include helpful content, passage ranking, neural matching, BERT, RankBrain and link analysis, and PageRank, it says, "continues to be part of our core ranking systems". Entities and E-E-A-T do not replace these basics, they help decide which source is chosen when several pages answer equally well.

How does this connect to GEO?

GEO, Generative Engine Optimization, is the work of being cited in the answers of AI engines. The mechanics are close to semantic SEO: an AI engine must recognise your brand as an entity, link it to a subject and judge the source reliable enough to quote. Our article on SEO vs GEO details the differences, and our complete GEO guide explains how to structure content so that it can be extracted.

Corroboration by third parties weighs heavily. In December 2025, Ahrefs analysed 75,000 brands and found that branded web mentions showed the strongest correlation with visibility in AI answers, while link metrics showed very weak correlations (Ahrefs itself reminds readers that correlation is not causation). The lesson: a coherent presence across the web counts more than a pile of links. For the local picture, see what AI engines actually cite in our study on the state of GEO in Belgium in 2026.

How to build topical authority, step by step

Here is the method we recommend to SMEs, in the order we apply it.

1. Choose one core subject. Pick the theme where you have real experience and a commercial goal, for example "Odoo for Belgian SMEs" rather than "digital".
2. Map the entities. List the people, products, places, standards and concepts around that subject (Odoo 20, Peppol, Belgian VAT, your founders) and note how they relate to each other.
3. Build the cluster. Plan one pillar page and six to ten supporting articles, each answering one precise question taken from real queries in Search Console.
4. Write answer-first. Open each section with a direct answer in two or three sentences, then develop it. This helps both featured snippets and AI extraction.
5. Link the cluster. Each supporting article points to the pillar and to two sibling articles, with descriptive anchor text.
6. Declare the entities. Show an author with a bio on every article, and add Organization, Person and Article structured data with sameAs links.
7. Earn corroboration and measure. Get listed and quoted on third-party sources, then track the cluster's impressions in Search Console and your citations in AI answers every quarter.

How do you set up entity optimization on your site?

Entity optimization means making your organisation, your people and your topics unambiguous for machines. It rests on two building blocks.

Organization: one identity, everywhere

Google recommends Organization markup with the name, url, logo and sameAs properties, and describes sameAs as the URLs of other pages with information about the organisation, such as social media or review-site profiles. Use exactly the same name, description and contact details on your site, LinkedIn, Google Business Profile and the directories where you are listed.

Person: authors who really exist

Google encourages a byline that leads to information about the author. In Article markup, the author should be a Person with a url or a sameAs that uniquely identifies them, for example a LinkedIn profile or a bio page. Google also warns that fabricated creator profiles, such as invented names or false credentials, are deceptive.

Why does duplicate content weaken topical authority?

Duplicate content is substantial content that appears on several URLs, on your site or elsewhere. Google groups very similar pages and chooses one version to show. The others are filtered, and your signals (links, mentions, clicks) are split between versions instead of building up on a single page. In a cluster, the worst case is two of your own articles answering the same question: they compete with each other.

Google rates its consolidation methods by strength: a redirect or a rel="canonical" link is a strong signal, a sitemap a weak one. Better still, merge the two articles into one stronger page and give every article its own question. On a multilingual or filtered site, also check that translated or parameter URLs do not create near-identical pages.

What mistakes should you avoid?

• Keywords without entities. Repeating a phrase never replaces naming the people, products and standards the subject involves.
• Anonymous content. An article without a named, verifiable author gives engines nothing to attach trust to.
• A scattered blog. Twenty articles on twenty subjects build no authority on any of them.
• An inconsistent identity. Different names, descriptions or details from one directory to the next make your entity blurry.

How does Nexuro apply this on its own blog?

We hold ourselves to the method above. Our blog is organised in clusters, each with a pillar and linked supporting articles, and we measure the effect with Search Console and a fixed panel of AI prompts.

• Clusters. Odoo 20 (new features, AI, website, accounting, CRM), Peppol e-invoicing, local SEO and GEO.
• An identified author. Every article carries a byline linked to a LinkedIn profile and to our about page.
• JSON-LD structured data. Article, FAQ and breadcrumb markup on every article, plus HowTo where it applies.

Frequently asked questions

What is semantic search in simple terms?

Semantic search means the engine tries to understand what you mean rather than only matching the words you type. Google uses AI systems such as BERT, RankBrain and neural matching to link a query to the concepts a page covers. A page can therefore answer a question without containing the exact phrase, provided it covers the subject precisely and completely.

What is entity optimization and where should you start?

Entity optimization is the work of making your company, your people and your subjects unambiguous for engines. Start with a single consistent identity: same name and description everywhere, Organization markup with sameAs links, named authors with a Person profile. Then map the entities around your core subject and cover them in a connected cluster of articles.

How do you know if you have topical authority?

There is no official score. Look at three signals: the share of your cluster's queries where you rank on the first two pages, the growth of impressions across the whole cluster in Search Console, and whether AI engines cite you on a fixed panel of prompts about your subject. Track them every quarter and compare with your direct competitors.

Is E-E-A-T a ranking factor?

No. Google states that E-E-A-T itself is not a specific ranking factor. Its systems use a mix of signals that help identify content with good E-E-A-T, and the quality raters who apply the guidelines do not control rankings. Treat E-E-A-T SEO as a checklist of credibility signals to make visible, not as a score to optimise.

Does duplicate content cause a Google penalty?

Google's documentation does not describe a penalty for ordinary duplicates. It groups similar pages and chooses one version to display. The real cost is diluted signals and the risk that the wrong page is shown. Fix it with a 301 redirect or a rel="canonical" link, or by merging the pages. Deliberate manipulation is another matter, covered by spam policies.

What influences Google ranking beyond keywords?

Google says its ranking relies on many automated systems that weigh many signals across a very large index. Among those it documents are helpful content, passage ranking, neural matching, BERT, RankBrain and link analysis, with PageRank still part of the core systems. In practice: useful content, clear structure, a credible author and links from other sites.

Key takeaways

• Engines now read meaning, not strings: write for questions and cover the related concepts.
• Make your company and your authors identifiable entities, with the same data everywhere.
• Topical authority is built with a pillar, linked supporting articles and a narrow, coherent subject.
• E-E-A-T is a framework for credibility, not a score, and trust is its most important component.
• Duplicate content splits your signals: one question, one page.
• GEO relies on the same foundations, reinforced by corroboration from third-party sources.

What next? Do Google and AI engines understand who you are and what you are an expert in? We can tell you in an audit of your entities, structured data, authors and cluster coverage. Request a free audit on our contact page or see our SEO service at work.

About the author

Timothy Jacqmin is co-founder of Nexuro Digital, a digital marketing agency and Official Odoo Partner in Belgium. He helps companies with their SEO, their Odoo upgrades and their visibility in AI engines. LinkedIn profile · About us

Sources

1. Google Search Central, Creating helpful, reliable, people-first content
2. Google, Search Quality Rater Guidelines, edition of 11 September 2025
3. Google Search Central Blog, Google Raters Guidelines and E-E-A-T, 15 December 2022
4. Google Search Central, A guide to Google Search ranking systems
5. Google, Introducing the Knowledge Graph: things, not strings, 16 May 2012
6. Google Search Central, Organization structured data and Article structured data
7. Google Search Central, Consolidate duplicate URLs and Link best practices for Google
8. Schema.org, sameAs property
9. Ahrefs, AI brand visibility correlations, 75,000 brands, December 2025