LLM citations is one of the most talked about topics in search, and somehow one of the least understood. There’s a lot of “tactics and tips” out there with little evidence, and most brands are still figuring out what any of it means for them in practice.
For Women’s History Month in March, we hosted a Sip & Search in London and invited Liv Day, SEO Lead at Digitaloft; Charlie Marchant, CEO at Exposure Ninja; and Rejoice Ojiaku, Senior Business Content Specialist at Wise and Founder of B-DigitalUK, to cut through it.
In this recap, you will get their takes on how LLMs surface content, what makes a citation worth having, and where to put your budget if you want visibility that lasts.
How LLMs decide what content to cite
1. What changes can a brand make in its content workflow to increase its chances of being cited by LLMs?
Rejoice Ojiaku, Senior Business Content Specialist at Wise:
Everyone is overcomplicating this. LLMs are not some mysterious black box that requires an entirely new discipline to crack. They are models that pick up clearly structured content and return it. That is the whole thing.
It’s worth knowing that you are no longer answering queries, but conversations. Write your content with that in mind, structure it so the model can read it without working hard, and you will get cited more often than you think.
LLMs favour digestible formats because they are easy to retrieve, like FAQs, clear headings, bullet points, and content that directly addresses the prompt. If the model has to dig to understand what you are saying, it moves on.
Domain authority still matters. So do the basics like meta titles and technical SEO, but at the content level, clarity is what moves the needle most.
2. What does query fan-out mode look like in practice, and how can we apply this concept to topical authority?
Liv Day, SEO Lead at Digitaloft:
Traditional search is simple. You type “best running shoes,” Google looks for pages that match that query or something semantically close, and surfaces a result. Query fan-out works differently.
When you type “best running shoes” into an AI system, it does not process just that one query. It splinters it into every possible sub-query underneath it, like:
- “Cheapest running shoes.”
- “Best for back pain.”
- “Best on a budget.”
- “Best black.”
- “Best with celebrity endorsements”
Then it pulls the answer that best satisfies all of them at once.
What that means for topical authority is that covering one angle of a topic is no longer enough. You need to be visible across every sub-query your audience might trigger.
The content earning the most citations in AI search already reflects this. A Glamour article I came across while shopping for a work bag had around 25 subheadings, each targeting a different variation of the same core query, and this is why it keeps getting cited across different prompts.
Your website content is not enough to win LLM citations
3. Walk us through your process for optimising existing content for AI search visibility.
Charlie Marchant, CEO at Exposure Ninja:
Before anything else, understand this: optimising your own content is a small portion of the picture. More than 85% of the time you get cited in ChatGPT, it comes from a third-party source. Your website is the starting point, not the whole strategy.
The process starts with prompt research. You need to understand which queries surface content in your niche, who is showing up, and how your brand is positioned relative to your competitors.
Are you coming up as premium? Budget? Or best value? This positioning matters, and once you know where you stand, you can start shaping it.
From there, you look at the pages on your site most likely to surface in AI answers.
For a SaaS brand, this is often the pricing page, because people are prompting LLMs to compare you against competitors on price, features, and benefits. Once you know which pages matter most, the work looks like this:
- Restructure those pages to surface the content you want cited
- Update your schema so LLMs are pulling what you intend
- Adjust headings and paragraphs to reflect the comparisons people are making
Then move to third-party sources. Identify which publications, review sites, and roundups are being cited for the queries you care about. If you are not in those articles, getting in front of the journalists and editors writing them becomes a core part of your visibility strategy.
4. Most AI citations now come from third-party sources like reviews, communities, and earned mentions. How much of search visibility is still about your website versus your wider online reputation?
Liv Day
Both still matter, but they serve different purposes.
On-site content handles the queries people ask directly about your brand, like opening hours, pricing, and specific product details. For those, the model pulls from your site, so keep that content accurate and well structured.
Off-site is where the weight moves. For objective queries, LLMs do not pull from your just website but also from your competitors, and heavily from third-party comparisons that sit between all of you.
About 40% of the content types cited by ChatGPT are product roundups and best-of lists, according to Ahrefs. If you are not in those roundups, you are invisible for those queries, regardless of how well your own pages are optimised.
That makes digital PR a direct LLM visibility play. You need to know which journalists are writing the roundups being cited, get your product in front of them, and make sure you are included. It is also worth researching across multiple LLMs rather than just one.
A mattress brand that Digitaloft works with gets cited by ChatGPT as the best weighted blanket based on a Guardian article. Type the same query into Copilot or Perplexity, and that article does not appear at all. Each platform cites different sources, so your strategy needs to account for all of them.
Not every citation is good for your brand
5. LLMs pull from all kinds of sources, but not every citation builds trust or preference. What makes a citation valuable from a content and brand POV?
Rejoice Ojiaku
Being cited does not automatically mean the citation is working for you. If it is not building authority, credibility, or driving traffic, it is just a mention. There are three things worth examining to determine whether a citation has any real value.
The first is where the citation is coming from. Just as domain authority matters in traditional SEO, source authority matters in LLM systems. A mention from a reputable publication carries more weight than one from a low-quality blog.
The second is context. Is your brand being positioned as a source of truth, or just one of several options? Order matters too. Our brains are wired to treat whatever appears first as the most credible, and that applies directly to how citations land in LLM responses. Being mentioned first signals authority, whether or not that is the model’s intention.
The third is narrative alignment. What is the model saying about you, and is it relevant to the prompt? LLMs hallucinate. They surface broken URLs and miss important context. A citation that misrepresents your brand or points somewhere it should not can do more harm than no citation at all.
6. AI models can sometimes misrepresent a brand, which hurts visibility and conversions. How do you spot misrepresentation in AI answers, and what can you do to fix it?
Charlie Marchant
Start by confirming whether you are being misrepresented. Larger brands tend to care most about this because they have specific positioning statements to protect.
A premium brand does not want to be described as cheap, nor does a specialist want to be framed as a generalist. Sentiment scores give you a starting point for understanding how LLMs are currently describing you.
Here are two examples that show how misrepresentation plays out in practice and how straightforward the fix can be:
- We ran a sentiment analysis for Beaches and Sandals and found significant negative sentiment. When we dug into the source, it was about grooms arriving at their honeymoon without a tuxedo and finding no on-site rental option. This frustration had spread across the web, and LLMs were reflecting it back.
- We also worked with a financial education client whose qualifications LLMs were consistently described as far more expensive than competitors, even though the pricing was identical. The issue was how their pricing structure was presented on their website.
We recommended updating the page to make the comparison clearer, and within three days, they were appearing at the top of LLM responses. Nothing changed in their actual pricing, only how clearly they communicated it.
Misrepresentation in LLMs is usually a content problem. Fix what the model is reading, and the output changes with it
The 3 metrics that show if your LLM strategy is working
7. AEO tools now track brand mentions and competitive positioning. Which metrics do you consider most valuable?
Rejoice Ojiaku
Many AEO tools are louder than useful. The metrics worth tracking are those that show influence, specifically how often you are appearing in LLM responses and where that visibility is coming from.
The share of LLMs is the first one worth tracking. Build a defined set of prompts relevant to your brand and monitor how consistently you appear in the answers. It gives you a picture of your standing relative to competitors, and not just a raw mention count.
Citation frequency and consistency are the second. Track how often you are mentioned and in which LLM systems. Tools like PromptWatch will tell you exactly which platforms are citing you and how often, so you can identify where you are strong and where you are losing ground.
The third, and most revealing, is source mixture. If the majority of your citations are coming from third parties, it tells you AI systems trust others to speak about your brand more than they trust your own content. If everything is coming from your own site, your third-party presence needs to be expanded.
The healthiest position is a balance between the two, where authoritative external sources are talking about you, and LLMs are also pulling directly from what you publish.
8. Moz research shows Google is increasingly citing its own ecosystems to keep users within its channels. How should brands respond when the platforms controlling discovery are also competing for attention?
Olabinjo Adeniran, Founder at Growth Case Studies:
YouTube is the most obvious example of Google favouring its own platforms. Brands that are not producing video content have a big problem.
Whether through creators, influencers, or a brand-owned channel, video gives you presence in an ecosystem Google is actively prioritising in its own responses.
Charlie Clark, CEO at Minty Digital:
The brands struggling most right now are those summarising what already exists. LLMs can generate that without citing anyone.
If you want to earn third-party coverage and stay visible in a zero-click world, produce something that did not exist before you made it. Original research and net-new knowledge are the two formats AI cannot surface without a source.
What to prioritise for long-term LLM visibility in 2026
9. Profound’s February 2026 report shows LinkedIn is now the number one cited domain for professional queries. How can brands use LinkedIn to get more LLM citations and improve brand authority?
Charlie Marchant
This data only matters if LinkedIn is showing up for the prompts you care about.
If your competitors’ LinkedIn content is appearing in ChatGPT or Copilot, go after it. If LinkedIn isn’t appearing in your brand’s queries at all, it is less of a priority right now.
For most B2B brands, though, it is worth pursuing.
Start by finding the right voices inside your company. Not everyone is comfortable posting on LinkedIn, and that is fine. Look for those who already have a natural inclination for it and build topical pillars around their role and expertise.
An HR lead’s natural territory is hiring, retention, staff benefits, and employee engagement. A marketing manager’s role is demand generation, brand, and campaign strategy. Start with what people already know and care about.
Once you have identified your topics and your people, get them to publish newsletters and long-form posts consistently on the themes you want to own. If someone is a better speaker than a writer, record a podcast or voice note first and build from that.
At Exposure Ninja, LinkedIn posts have influenced both LLM citations and Google AI Overviews, and the two are more connected than most people realise.
10. Social media is a major source of citations. How should brands show up if they want to be cited without sounding forced or overly polished?
Hannah Smith, Head of Content at Women in Tech SEO:
[Image: Photo of Hannah Smith at the event]
Partnerships are one of the most effective approaches here. In fashion or beauty, that might look like an influencer relationship where someone your audience already trusts creates the content and carries the message. People are far more likely to engage with that than with a brand talking about itself.
For B2B, the same principle applies. Pair your brand with people in a relevant community and have them create content together.
Erin Simmons, Managing Director at Women in Tech SEO:
[Image: Photo of Erin Simmons at the event]
Put your audience in front of your brand, not the other way around.
- Feature your customers
- Amplify their knowledge and their skills
- Make them the centre of the content rather than a faceless organisation
When the content is about the people you serve rather than the company doing the serving, it feels human because it is.
11. If you had extra budget in 2026, where would you invest it for long-term LLM visibility?
Emma-Jane Stogdon, Organic Content Manager at Wise:
Invest in offline visibility. The brands winning with LLMs are the ones people talk about outside of search. Events, PR, industry conferences, and activations help, but none of that works if your teams are still operating in silos. You need a fully integrated strategy where brand, SEO, PR, and social are all working towards the same goal at the same time.
Adewale Adetona, Senior Product Marketing Manager at Arqiva:
Invest in partnerships. We published white papers for two products in partnership with three customers, and the results exceeded anything we would have achieved alone. Customer voice drove visibility, created content third parties wanted to reference, and broke down barriers for our sales team in ways that internal content never had.
Ryan Glass, Performance Experience Lead at Performics:
Look at your top-performing content from the last 24 months. That is what your audience has already told you they want from your brand. Build the information-gain version of it and distribute it across every format available to you, video, blog, social, and podcast.
Ross Simmonds has a framework worth looking up called “create once, distribute forever.” Repetition through distribution is how you build a consistent brand signal that LLMs keep picking up over time.
Charlie Marchant
Invest in digital PR and citation building. Brands that consistently appear in LLM responses have spent years building clear, consistent positioning before LLMs were even part of the conversation.
The Ordinary is a good example. Two proposition statements, best-value skincare and science-backed skincare, pushed consistently across Cosmo, Glamour, and everywhere they could earn coverage. By the time ChatGPT arrived, the work was already done.
Start that work now, and it compounds. Keep doing it, and others start repeating your message for you. Influence what people say about you, put out the message you want associated with your brand, and you will find it becomes a flywheel.
Rejoice Ojiaku
Invest in training. Not deep technical training, but enough for every team to understand what the others are doing. Your social team should understand SEO. The PR team should understand how LLMs work. Answer engines pull signals from multiple channels, and if your teams are not talking to each other, no amount of optimisation will close that gap. Marketing only works as an ecosystem when everyone inside it is aligned.
Conclusion: Get your house in order before chasing LLM visibility
AI citations do not come from tweaking a page or stuffing in more FAQs; you need to build your brand in a way that is clear, credible, and consistent enough for an AI to trust.
Sort that out first, and the visibility becomes easier.
Johnson Ishola
Johnson Ayomide Ishola is a content marketer at the FCDC. With a background in Economics, he brings a strategic, analytical edge to storytelling and brand development.
He specializes in lead generation and community-driven content, focusing on building authentic audience connections. At the FCDC, Johnson develops initiatives across various events and programs, using data-backed insights to craft narratives that drive engagement and growth.
