Every so often someone asks me if we do “entity SEO,” with the same tone you’d use to ask if a restaurant has vegan options. As if it were a separate item on the menu, a kind of peculiar side dish. My short answer is yes. And no.
This post is the long answer.
Every few years SEO gets a new name. “Entity SEO” is one of the ones that stuck around longest, and now it’s back, riding on AI’s coattails. It’s worth separating what’s real from what is, bluntly: a hoax.
What entities are: Google’s Knowledge Graph
An entity is something that can be identified without ambiguity: a person, a company, a place, a concept. Entity SEO (or semantic SEO) is the work of helping search engines and AI models recognize those entities in your content and understand how they relate to one another.
On May 16, 2012, Google introduced the Knowledge Graph with a post by Amit Singhal titled “things, not strings”. The idea was for the search engine to stop matching words and start understanding what actual thing in the world they referred to. At launch it had more than 500 million objects and 3.5 billion facts about them.
That’s why Google distinguishes between the Taj Mahal the monument and Taj Mahal the musician, or between Apple the company and the apple the fruit. In May 2020, Google reported that the graph already had more than 500 billion facts about 5 billion entities (Google).
So the question in the headline has an easy first answer: entities exist, and Google has been using them for more than a decade. Let’s keep going:
So, what is entity SEO or semantic SEO?
This is where the debate starts. If Google already runs on entities, then all the SEO you do happens inside a system built on entities. Ahrefs summed it up in a post titled “Entity SEO is just SEO”. In that post, Si Quan Ong quotes his colleague Patrick Stox: identifying entities is more Google’s job than ours.
The same post shows that the tactics usually sold as entity SEO are familiar ones: getting a Wikipedia page, setting up a Google Business profile, linking internal content well. You’d do all of that anyway, label or no label, because they’re part of the web of SEO best practices we’ve built over more than two decades. Google did the work, and all of those signals are, precisely, part of the indexes and the information that ends up being scraped and returned to the user in the form of Knowledge Graphs… or AIOs, right?
I mostly agree with Ahrefs, with one caveat. Thinking in terms of entities does change one thing in my workflow: the question I ask myself before writing. With keywords, I ask “what are people searching for?” With entities, I ask “what things does this text need to talk about so a machine understands what it’s about?” The second question produces better writing, whatever you call the discipline. Especially these days.
Three myths about entities and SEO
Myth 1: “Schema markup makes you rank better”
Google has denied this many times. Danny Sullivan, Google’s Search Liaison, said it plainly: “schema doesn’t make you rank better” (Search Engine Roundtable). John Mueller had already said back in 2018 that there’s no generic “boost” from using structured data (Search Engine Journal). What schema does do is enable rich results and help describe your content.
Myth 2: “You have to optimize your entities”
Entities are identified by Google. Your job is to make sure your content is clear about what it’s about and that your brand is described consistently everywhere.
Myth 3: “You need an entity tool”
It can help with auditing, for example using Google’s Natural Language API. It doesn’t replace writing well about a topic or getting mentioned on sites that matter. Please, I’m begging you, stop buying licensed software!
Schema is the clearest example of this. It’s one of the first things added in an audit because it’s cheap and tidy, and also one of the first things to disappoint when someone expects to see rankings jump the following month. That said, implementing it is practically automatic, and skipping it is just careless. Personally, I’ve seen it work. But you already know how this goes.
What actually did change with AI
There’s one place where thinking about entities became more useful: LLMs. For ChatGPT, Gemini, or AI Mode to recommend your brand, they first need to know who you are, what you do, and how your brand differs from another one with a similar name. It’s the same logic we explained in what agentic search is: the agent chooses between brands it understands.
AI visibility tools already reflect this. Ahrefs Brand Radar’s help docs warn that AI share of voice can shift a lot depending on how you set up the entities for your brand and your competitors (we compared these in our guide to AI visibility tools). And OpenAI already lets you claim your company’s profile inside ChatGPT, which is, deep down, an entity card (we covered that here).
In practice, entity work for AI looks a lot like entity work has always looked:
- One name, one description. Make sure your brand is described the same way on your site, LinkedIn, directories, and profiles.
- Organization structured data with your official profiles linked (
sameAs). - Third-party mentions in media and industry sites, which is where models learn who you are.
- Content that covers the full topic, including the related concepts an expert would mention.
- Wikidata or Wikipedia if you meet the notability criteria. If you don’t, don’t force it.
In Spanish, this comes with an extra problem. Many brands in the region are named after a common word (a fruit, a verb, a neighborhood), and for a machine that’s pure ambiguity. If your brand has that problem, the consistency from point 1 stops being a detail and becomes the foundation for everything else. If you’re interested in how visibility is approached beyond Google, you can read our interview with Sebastián Galanternik on search everywhere optimization.
Our verdict
Entities exist. Entity SEO as a separate discipline, with its own tools and its own budget, is largely a new way of selling SEO done right. So it’s still worth asking, when thinking about your SEO: can a machine that doesn’t know you understand who you are from what’s published about you?
At Postdigitalist we don’t sell entity SEO as a separate service, and if someone sells it to you that way, ask them to show you what they actually do differently. What we do is make sure your brand is coherent and offers solid content aimed at the conversions you’re after. For an LLM, that thematic coherence is called an entity.

By Circe Galanternik
Frequently asked questions
What is an entity in SEO?
It’s something that can be identified without ambiguity, such as a person, a company, a place, or a concept. Google organizes them in its Knowledge Graph and tracks how they relate to one another.
What’s the difference between entity SEO and semantic SEO?
They’re used almost as synonyms. Both aim to get the search engine to understand the meaning and context of the content, beyond the exact words.
Does structured data improve rankings?
According to Google, not directly. It helps enable rich results and describe content in a way that’s legible to machines.
How does my brand show up in the Knowledge Graph?
There’s no form to fill out to get in. It helps to have consistent information on your site and profiles, organization structured data, and mentions in trustworthy sources. If Google shows a panel for your brand, you can claim it.
