The Complete
AI Citation
Playbook

Let's start with the big picture, because everything else in this guide flows from it. For twenty years, "being found on Google" meant appearing in a list of ten blue links when someone typed a query. The game was ranking. You got a keyword, you built content around it, you got backlinks, and you climbed the list. The person clicked your link. You won.

That model is not broken. But it is being fundamentally restructured. Google's AI Overviews now appear in over 60% of all searches. That means in more than half of all queries, the person sees a synthesised AI-written answer at the very top of the page before they ever reach the blue links. In many cases, particularly for informational and commercial queries, they get the answer they were looking for right there, and they do not click anything.

The January 2026 Gemini 3 switch is worth sitting with for a moment. Forty-two percent of previously cited domains got reshuffled overnight. Brands that had worked hard to appear in AI Overviews found themselves absent the next morning. Others who hadn't optimised specifically for AI suddenly showed up. This tells you something important: the AI answer layer is not stable. It is actively being improved, retrained, and reconfigured. Your best insurance is depth of real authority, not clever one-time optimisation tricks.

The practical consequence for your business is that you now need a two-track visibility strategy. Track one is the traditional SEO you already know: ranking for your target keywords in the organic results. Track two is the new AEO and GEO work: ensuring that when AI engines synthesise answers in your space, they have your brand, your facts, and your credibility to draw from. The two tracks reinforce each other, but they are not the same thing.

The scale of what is happening in AI search is easy to underestimate because it crept up gradually and then arrived all at once. Perplexity alone processes over 780 million monthly queries and that number is growing fast. AI search visits grew 42.8% year on year from Q1 2025 to Q1 2026. Gartner projected in early 2024 that traditional search engine volume would drop 25% by 2026, with AI chatbots taking a substantial share of that intent. Whether Gartner's exact number proves accurate or not, the directional trend is unmistakable to anyone paying attention.

Here is the thing that makes this urgent rather than just interesting. The brands establishing AI citation presence right now are building compounding advantages. An AI engine that has seen your brand consistently cited across Wikipedia, LinkedIn, Reddit, YouTube, and trusted industry publications is going to keep citing it. That citation pattern feeds back into training data over time. Early movers in AEO and GEO are building authority that will take latecomers years to replicate.

The good news, and the Princeton GEO research makes this explicit, is that GEO is a genuine levelling opportunity. Their study found that adding statistics and genuine citations to content improved AI visibility by up to 115% for sites ranked fifth or lower in the SERP. The number one site actually saw little improvement because it was already being cited. If you are not currently dominant in Google, the AI answer layer gives you a real chance to leapfrog competitors who have more backlinks but less clarity, less structure, and less third-party corroboration.

Think of the major AI answer engines as journalists working for different publications with different editorial standards. Perplexity is your hyperactive breaking-news reporter who cites 21 sources per story and updates everything in real time. ChatGPT Search is more like a thoughtful feature writer who leans on Wikipedia and established publications. Gemini is the in-house Google expert who can pull from Maps, Business Profile, and a decade of search ranking data. Claude is your careful academic who wants primary sources and properly attributed facts. Understanding who you are pitching to changes everything about how you write and where you place content.

The key takeaway from comparing these six engines is that a strategy which depends entirely on your own website will leave you underserved everywhere except Gemini. The research consistently shows that third-party mentions, community presence, and off-site authority are the dominant forces. Your website is the credibility anchor — the source that everything else points back to — but the weight of your AI visibility is carried by what the wider web says about you.

Here is a number that tends to stop people mid-coffee: BrightEdge found that YouTube is cited 200 times more than any other video platform across AI engines. That is not 200 percent more. That is 200 times more. The platform with the second highest video citation rate is not even in the same conversation.

But it gets more interesting than that. Ahrefs found that YouTube mentions correlate with AI visibility at a coefficient of 0.737. To put that in context, this is the single strongest statistical predictor of AI citation that researchers identified across all the major 2025 and 2026 studies. Stronger than backlinks (0.218). Stronger than brand mentions (0.664). Stronger than domain rating. When the data is saying this loudly, you pay attention.

For local businesses specifically, YouTube creates an almost unfair advantage over competitors who are not using it. A plumber in Leeds who films a 10-minute video explaining common boiler problems and how to diagnose them before calling an engineer has created something that Gemini will cite when someone in Leeds asks that question. Their competitor, who wrote a 300-word blog post about the same topic with no video, will not get cited. The content quality is similar. The format difference is enormous.

Grok follows a similar pattern. In Ahrefs' June 2026 analysis of 1.9 million queries, YouTube was the second most cited domain on Grok at 15.1%, just behind Reddit at 16.3%. X/Twitter itself only accounted for 1.4% of Grok's formal citations, which tells you something interesting: Grok uses X for real-time sentiment and context, but it cites YouTube and Reddit for factual substance. The lesson is to be present on X as a signal but invest your content energy in YouTube as a citation source.

There is a useful analogy here. Your website in the age of AI citations is like your LinkedIn profile in the age of professional networking. It is not where the relationship starts (that happens elsewhere), but if someone finds out about you and goes to check you out, a weak or confusing profile kills the deal before it begins. Your site needs to be the clearest, most credible, most well-structured version of what you do and why you are trustworthy. Then the rest of your presence points people (and AI) back to it.

Schema is not magic. Ahrefs ran a controlled study of 1,885 pages adding schema and found that citations "barely moved" for pages already in the consideration set. But schema is still worth doing because it helps AI engines understand your entity clearly, which is different from directly causing citations. Think of it as making sure the AI is not confused about who you are, what you do, and where you are based. Confusion does not get cited.

The research on this point is not subtle. Brands are 6.5 times more likely to be cited through third-party sources than through their own domain. A study of over 21,000 brand mentions across ChatGPT, Claude, and Perplexity found approximately 85% came from external validation rather than the brand's own site. This is the fundamental shift in how visibility works. You are no longer just publishing content. You are building a reputation ecosystem where dozens of trusted sources all independently confirm that you are who you say you are.

There is a loud debate in the SEO community about whether links matter for AI citations. The answer is yes and no, and both camps are partly right. Ahrefs found that backlinks correlate with AI visibility at r = 0.218, while brand mentions correlate at r = 0.664. Which means brand mentions are roughly three times as predictive as backlinks when it comes to whether AI engines cite you. A different study found that domain authority explains under 4% of citations. Yet SE Ranking's analysis of 129,000 domains found that referring domains were the single strongest predictor of ChatGPT citations, and sites with over 32,000 referring domains were cited 3.5 times more than those without.

The reconciliation is this: you need a baseline level of link authority to be crawled, indexed, and taken seriously as a domain. Around DR 30 or more appears to be the threshold where you enter the AI consideration set at all. Below that, you may not even be in the running. But above that threshold, accumulating more links does not reliably buy you more AI citations. What buys citations is content structure, freshness, brand mentions, and third-party corroboration. Links build the foundation; everything else builds the house.

Local businesses often ask whether mentioning their specific town, city, or region actually influences whether AI engines cite them for local queries. The answer is clearly yes, but the mechanism is more interesting than just stuffing a city name into your content. AI engines doing local query synthesis are trying to triangulate your location, relevance, and authority from multiple data sources at once. The more consistent and rich those signals are, the more confidently an engine can include you in a locally-anchored answer.

Not every business starts from the same place. Use these diagnostic triggers to prioritise where to focus your effort first, because trying to do everything simultaneously usually results in nothing getting done properly.