You’ve been publishing. The blog is active, the posts are solid, and your rankings haven’t collapsed. But you searched for your brand in Perplexity last week. ChatGPT doesn’t mention you. Google’s AI Overviews cite three of your competitors and not once pull from your site.
The usual advice? Publish more. Structure your content better. Add FAQs.
That advice is wrong — or at least, it’s aimed at the wrong problem.
Most businesses that are invisible to AI answer engines don’t have a content problem. They have an entity problem. And until that’s fixed, no amount of publishing will move the needle.
Here’s what that means, why it matters, and what to do about it.
How AI engines decide who to cite
When you search on Google, the engine matches your query to documents. Ranking is a relevance competition — the page that best answers the query wins the placement.
AI answer engines work differently. Before ChatGPT, Perplexity, or Google’s AI Overviews will cite a source, they run a prior check: is this a real, verified entity?
An entity, in search terms, is a disambiguated thing — a person, brand, organisation, or concept that the knowledge graph can identify with confidence. Google, OpenAI, and Perplexity all draw on interconnected knowledge graphs to resolve whether a source is who it says it is before recommending it.
Think of it like a building permit. Your content is the structure. Your entity footprint is the permit. Without the permit on file, the building doesn’t officially exist.
If your entity is ambiguous — inconsistent name across the web, no structured data, no corroborating third-party mentions — the AI treats you as unverified. It will pull from a competitor whose entity it can resolve confidently, even if your content is stronger.
This is the misdiagnosis most businesses are living with: they’re optimising the structure while the permit is missing.
What a weak entity footprint actually looks like
Entity problems are invisible until you know what to look for. Here are the most common signals:
- Your business name is inconsistent across the web — the website says one thing, Google Business Profile says another, LinkedIn uses a variation, and your social bios don’t match any of them.
- There’s no structured data on your site. No Schema markup identifying your organisation, the services you offer, or the person behind the brand.
- Your social profiles exist but aren’t linked from the website — and the website isn’t linked from the profiles. The graph has no connective tissue.
- You have no editorial mentions. Directories and press releases don’t count. Third-party publications that name your brand in context — that’s what builds corroboration.
- There’s no Google Knowledge Panel for your brand name. This is the most visible signal that your entity hasn’t been established in Google’s graph, which is the same graph powering AI Overviews.
A brand doesn’t need all of these resolved overnight. But an AI engine needs enough cross-referenced signals to confirm you are who you say you are. Thin signals mean you get skipped.
The three-layer entity stack
Entity authority isn’t a single switch. It’s a stack — three layers that together tell AI engines your brand is real, consistent, and active. Here’s how to think about each one.
Layer 1: Identity consistency
This is the foundation. Every surface where your brand appears — website, Google Business Profile, LinkedIn, social profiles, directory listings — needs to use the same name, URL, and description. Not similar. The same.
On the technical side, this layer includes Organisation or Person Schema markup on your site, with sameAs properties linking out to every verified profile. The sameAs link is what tells the knowledge graph that all these surfaces are the same entity.
Layer 2: Corroboration signals
Identity alone isn’t enough. The graph needs third-party confirmation. This is where most brands have the biggest gap.
Corroboration comes from editorial brand mentions — articles, interviews, podcast appearances, or resource roundups that name you in context. It also comes from a LinkedIn presence with accurate role and affiliation data, and ideally a Wikidata entry. Wikidata is accessible to any brand, not just large enterprises, and it’s one of the primary sources AI engines use to resolve entities.
Layer 3: Freshness and activity
Entity authority isn’t a one-time setup. AI engines treat it like a loop — they check, cite, and re-check. Brands that maintain active, consistent signals compound their citation frequency over time. Brands that set it up once and go quiet decay.
Freshness signals include recent content that names your brand in context, active social profiles with consistent bios, and Schema markup that hasn’t gone stale. An outdated schema is a quiet trust penalty.
The compounding effect
Every time an AI engine successfully resolves your entity and cites you, it reinforces the signal for the next query. Citation is not linear — it compounds. That’s why getting the infrastructure right early matters more than publishing volume.
Three checks you can run today
No tools required. These three checks will tell you quickly whether your entity is in good shape or needs work.
- Search your brand name in ChatGPT and Perplexity. Does the AI know who you are? Does it describe you accurately, or does it hedge, confuse you with someone else, or come back blank? Vague or wrong answers mean your entity is unresolved.
- Google your brand name and look for a Knowledge Panel — the card that appears on the right side of results with your logo, description, and links. No Knowledge Panel means your entity hasn’t been established in Google’s graph. AI Overviews run on that same graph.
- Take three sentences from your About page and paste them into Perplexity. Ask: who said this? If Perplexity can’t attribute the content back to you, your content is not connected to your entity in the knowledge graph.
If any of these checks came back with a problem, you’ve found the real gap. The question now is how to close it systematically.
Citation is a system, not a content problem
The shift from search engine to answer engine changes the game in a specific way: it moves the primary gatekeeping mechanism from relevance to trust. You can have the most relevant content on the web, but if the AI can’t verify your entity, it will cite someone it can verify instead.
That’s not a content problem. It’s an infrastructure problem — and infrastructure problems have infrastructure solutions.
The entity stack — identity consistency, corroboration signals, freshness and activity — is the foundation of every site I audit before touching a single piece of content. Because until the permit is in place, the building doesn’t officially exist.
Ready to audit your entity footprint?
If you ran the three checks above and didn’t like what you found, that’s exactly where we start. Not with a content calendar. With a system.
Book a free 30-minute audit. We start with entity, not keywords — because that’s where the gap usually is.