SEO is changing forever. Instead of traditional SERPs with links to websites, we are moving to citations in AI-generated answers.
This means that all website owners, SEOs, and marketers need to understand AI citation mechanics. By understanding this, they can increase their chances of their website appearing in AI answers.
What Are AI Citation Mechanics
AI citations mechanics refer to how AI search engines, like Google’s AI Overviews, ChatGPT, and Perplexity, retrieve, choose, and attribute sources when generating answers.
Backlinks Vs AI Citations
| Aspect | Backlinks | AI Citations |
| What They Are | Links from other sites that boost rankings. | AI search engines referencing your content in answers. |
| Visibility | Mostly hidden; users don’t usually see them. | Front-facing; users see the source listed. |
| Trust Impact | Indirect through better rankings. | Direct by showing trusted sources in answers. |
| Selection Factors | Link quality, quantity, anchor text. | Content relevance, originality, and formatting. |
| Examples | Blog links to your site. | Google AI Overview citing your page. |
| SEO Focus | Earning backlinks from strong sites. | Structuring content to be AI-citable. |
| Effect | Improves rankings over time. | Boosts immediate visibility and credibility |
How AI Chooses What to Cite
AI doesn’t “rank” pages like traditional SEO. Instead of ranking an entire page, they focus on extracting specific passages that best answer the user’s question. Because of this, individual paragraphs or sections often matter more than the page as a whole.
When AI scans your content, it’ll scan the entire page. However, it’s looking for concise, self-contained passages that answer the query directly. If part of your content provides a clear, relevant answer, the AI could cite your page, even if it isn’t ranking well on the SERPs.
Factors Influencing Which Content AI Cites
- Passage-Level Relevance: AI focuses on passage rankings, so it priorities on the content that directly answers the user’s question rather than the entire content on the page.
- Semantic and Contextual Understanding: AI focuses on meaning, favouring content that has semantic relevance to the user’s intent.
- Structured Content: Like humans, AI needs to read content easily, so it needs to be appropriately structured using headings, schema, image alt tags, etc.
- Fact-Based and Actionable Information: Pages rich in unique facts, data, specifications, instructions, and comparisons generally get cited more.
Content Types That Get Cited A LOT
XFunnel performed a 12-week study, analysing over 768k AI citations and discovered that product-related content dominated.
In total, content relating to products, like “best of” articles, comparisons, product pages, etc., accounted for 46% to 70% of all cited sources.
This was then followed by News and Research content, 5-16%, affiliate content, single digits to 20%, user reviews, 3-10%, blogs, 3-6%, and PR content, under 2%.
Based on this research, it’s clear that the AI chooses content that is unique, rich in facts and instructional detail, and is structured in a way that the AI can easily parse.
RAG and the Role of Generative AI
RAG, or Retrieval-Augmented Generation, is somewhat the “magic” behind AI citation mechanics. Put simply, it’s how AI search engines retrieve data and then generate an answer.
This technique follows a hybrid approach. First, the AI retrieves relevant data from knowledge sources. In most cases, web content. Second, the LLM uses this data to create an appropriate answer.
As a result, the quality of the AI’s response relies heavily on the data it’s fed. If it doesn’t extract from quality and trusted data sources, the output won’t satisfy the user. This is why citations matter A LOT with AI search engines.
Why Citations Matter in AI Search
Generally, AI search engines are trained on several knowledge sources. These mainly include web content and a private or public database.
If the data fed to the AI is poor, the generated output will reflect that. Even worse, when it doesn’t have relevant information to provide an answer, it can experience hallucinations.
Hallucinations are when an AI “makes up” the output. It doesn’t know the query’s answer, so it just makes something up. This could be made up content, links to references, etc.
AI can provide incorrect answers. This is why citations will remain important for users. It gives them a chance to verify whether the information provided is factual, trusted, and true.
How to Optimise for AI Citations
- Use Schema Markup
AI search engines need to understand your content in a language they understand. This is why you need to use schema markup.
Schema markup is a form of structured data, a language, so to speak, that non-human entities understand to become aware of the meaning behind a page and element.
For example, a popular type of schema, a HowTo schema, is used on how-to guides. It clearly suggests to the AI, search engine, etc., that this page is a how-to guide and therefore should show in how-to guide-like search queries.
Simply, using schema markup makes it easier for AI search engines to understand the content on your web page. And, the easier it is to understand, the higher the chances it’ll appear as a citation.
- Apply Semantic HTML
Semantic HTML are elements like <article>, <section>, <header>, <h1>–<h6>, <p>, and <nav>. They clearly define the structure and hierarchy of the content, which, again, is easily understood by non-human entities.
Similar to using schema markup, using semantic HTML makes your content more machine-readable. As a result, it increases your chances of getting cited for a user’s search query.
- Prioritise Clarity
When creating content, focus on clarity throughout. This means writing clear, concise sentences as well as correctly labelling any sources that you mention within your content.
Just think clarity, clarity, clarity. Be so clear in your writing (using bullet points, images, etc.), referencing, and external linking that humans and AI search engines can easily interpret it.
For example, when referencing a study in your content, don’t write “15% increase” and then external link the text, say, “According to a 2025 study by Amazon, they saw a 15% increase…”, linking the first part of the sentence.
By being clear, you help both parties, AI search engines and humans, better understand your content and its context. By doing this, you massively increase your chances of being mentioned as a citation in a search query.
- Focus on Internal Linking
Internal links help AI search engines understand the legitimacy and topical authority of your website.
Again, by using clear anchor text as well as good internal linking practices it allows them to understand the purpose of your website and can increase the chances of citations.
From Rankings to Citations: SEO 2.0
SEO today is changing, and there’s nothing stopping that. Instead of chasing rankings, websites will move to chasing citations.
This new type of SEO is actually called AEO, otherwise known as answer engine optimisation. Instead of using SEO to rank #1 in the SERPs, marketers will use AEO to increase their chances of appearing in AI answers and their citations.
And, due to AI interfaces providing entire answer outputs, this means an increase in zero-click SEO. In fact, research from Sparktoro suggests that 58.5% of searches made by Americans in 2024 resulted in zero clicks.
This, as you can imagine, will result in less traffic. Therefore, the new type of traffic will be brand discoverability – how often your brand appears in citations as a trusted source.
The maths is simple. If your brand continuously appears in AI answers, people will, over time, trust you more. As more people trust you, you’ll become a go-to source of information, and, most likely, business.
Final Thoughts and Next Steps
Without question, AI citation mechanics are changing how we approach SEO. Rather than ranking in the SERPs, ranking in AI answers is going to be the next go-to move.
Because of this, former SEOs need to turn into AEOs, creating content for AI-friendliness by using schema and clear semantics and creating helpful content.
This doesn’t mean forgetting about SEO, though. Traffic still comes from the SERPs, and several SEO practices, like E-E-A-T, can increase the trustworthiness of your content, which, ultimately, can increase your chance of getting a citation.
What I recommend for anyone is to start with the SEO basics. However, focus more on schema, clear content (using semantics), and structure your content correctly, for example, using heading tags.
Therefore, go through your website, audit your content to see how you can improve its machine readability, and measure the outcome of your changes.
