LLM citations are now an important SEO metric. But unlike traditional search, there's no playbook, keywords, or algorithm update to reverse-engineer.

For our Women's History Month edition of Sip & Search SEO Meetup, I invited Liv Day, SEO Lead at Digitaloft; Charlie Marchant, CEO of Exposure Ninja, and Rejoice Ojiaku, Senior Content Specialist at Wise, to share tips on getting LLM citations.

In this recap, you'll get expert insights on how AI models retrieve information, what makes a great citation, and where to put your budget if you want long-term LLM visibility.

1. What changes can a brand make in its content to increase its chances of LLM citations?

Rejoice Ojiaku

LLM systems are not as complex as everyone is making them out to be. The sensationalizing needs to stop, because the answer is simpler than most people want to admit. 

Instead of answering queries or questions, you’re now answering conversations. Keep this in mind when creating content, and you will get cited more often than you think.

These are not search engines with an index, but models layered with retrieval-augmented generation. 

LLMs love digestible content formats like FAQs, listicles, and bullet points because they’re easy to retrieve. If your content is clearly structured and directly relevant to the prompt, it picks it up and surfaces it. If the model has to work to understand what you are saying, it moves on.

Jamie Indigo said it better in this quote below:

Source: The Dark Side of AI No One Talks About

However, Domain Authority still matters, as do the fundamentals like meta titles and technical SEO. But with LLMs, the most important thing you can do is make your content easier to read because clarity outranks almost everything else.

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, Digitaloft:

Query fan-out is how AI systems like LLMs and Google's AI Mode retrieve and present information. 

Traditional search is straightforward; you type best running shoes, and Google scans its index for pages that match the query or something semantically similar.

Query fan-out takes that same search and splinters it into multiple sub-queries, like: 

• Cheapest running shoes
• Best running shoes for back pain
• Best running shoes on a budget
• Best black running shoes
• Best running shoes with celebrity endorsements

Then, it presents what it determines as the best answer across all of them. 

I encourage you to read this piece by Pete Meyers, which explains how query fan-out works and its impact on SEO.

The implication for topical authority is that you’re not optimizing for a single keyword. You need visibility across all sub-queries that could branch from your core topic.

The content earning citations in AI search already reflects this. For example, I was recently shopping for a work bag and came across a Glamour article with around 21 subheadings, optimized for queries like best vegan leather tote bag and Best Leather Shoulder Bag For Work. That level of sub-query coverage is why Glamour is cited in different prompts.

3. Walk us through your process for optimizing existing content for AI search visibility

Charlie Marchant, CEO, Exposure Ninja:

According to AirOps research, 85% of AI citations come from third-party sources. Your on-site content matters, but it’s one piece of a larger pie.

Start by identifying which pages on your site are most likely to surface in AI answers. If you work in SaaS, people are comparing you to competitors through entities like pricing, benefits, and features. That means your pricing page is probably more important for LLM visibility than your blog.

Next, execute these steps:

Prompt research

Analyze what content surfaces for the queries you care about. You can do this manually in a spreadsheet, but Moz has a prompt research tool to speed it up. 

Log in to your Moz Pro account, hover your mouse over AI Research, and click Prompt Suggestions. Enter a topic relevant to your brand and click Analyze.

You’ll get a list of prompt suggestions you can track in Moz AI Visibility. The goal is to understand what's being cited and where you're showing up.

Crystal Carter has a detailed guide on building a target prompt list and using the data to drive impact.

Competitive analysis

Compare how you appear against your competitors. Understand whether you're coming across as premium, cheap, or best value, and if the positioning aligns with your intent. Next, identify which parts of your content LLMs reference and whether those are the parts you want cited.

Page structure

Once you know what needs to change, restructure the page so it surfaces the content you want. That means rewriting paragraphs, updating headings, and refining the questions you answer. It also means updating schema so LLMs surface what you want them to.

John Iwuozor has an excellent guide on optimizing content for AI Mode that would also apply to other LLMs.

4. What makes a citation valuable from a content and brand POV?

Rejoice Ojiaku, Senior Content Specialist, Wise:

Getting cited doesn't automatically make a citation valuable. If it's not building authority, credibility, or driving traffic, then it’s not working for you.

There are three ways to evaluate a valuable citation:

Citation source

Just as Domain Authority matters in traditional SEO, the authority of the citation source matters in LLM systems. A mention from a reputable publication carries more weight than one from a low-quality blog.

The context of the citation

Is your brand being positioned as a source of truth, or just another option? Citation depth matters because our brains are wired to treat whatever appears first as the most important. 

This applies to citations in LLM responses, where being mentioned first signals authority, whether or not that's the model's intention.

Moz AI visibility tool offers a great way to track citation depth in AI search platforms like ChatGPT and Gemini.

For example, in the screenshot above, you can see that Toyota and Nissan are often mentioned near the top, while Honda and Mazda are not. On the right, you can compare historical performance over time to understand what’s working and what's not.

Here’s a detailed workflow to set up your AI visibility dashboard in Moz

Narrative alignment

Consider what the model is saying about you and the relevance to the prompt. LLMs hallucinate and miss out on important context. A citation that misrepresents your brand or links to the wrong page could do more harm than good.

Remember, citations are only as valuable as the authority behind them, the citation depth, and the accuracy of the narrative around it.

5. Most AI citations now come from third-party sources. How much of search visibility is still about your website versus your wider online reputation? 

Liv Day

Both are still very important, but they serve different purposes.

On-site content handles brand queries. If someone searches for your opening hours or pricing, the model pulls the information from your site, so make sure your content is accurate and well-structured.

Off-site is becoming more important. For objective queries like "HubSpot vs Pipedrive," LLMs look at information from both brands but lean heavily on third-party comparisons. 

Prioritize digital PR and reverse-engineer the prompts you want to show up for. Then, test those prompts, look at what they’re citing, and make sure you're visible there.

According to Wix, 52% of AI citations come from listicles, articles, and product pages.

 So, if The Guardian or a major media brand publishes a roundup of your product category and you're not in it, you may not get visibility in AI search. In this scenario, the fix is targeted outreach to the journalists writing those roundups.

It also pays to research across LLMs, not just one. ChatGPT has partnerships with specific websites and organizations that influence what it cites, as evidenced by this Ziff Davis study from February 2025. 

At Digitaloft, we work with a mattress brand that ChatGPT cited as the best weighted blanket, based on a Guardian article. Type the same query into Copilot or Perplexity, and the article doesn't appear because each platform cites entirely different sources.

6. AEO tools now track brand mentions and competitive positioning. Which metrics do you consider most valuable? 

Rejoice Ojiaku

Focus on metrics that measure influence, specifically traffic and visibility from answer engines.

Below are three metrics that tell this story far better than anything else:

Share of LLM

Use a tool like Moz Prompt Suggestions to build a prompt list relevant to your brand and track those prompts in your AI visibility dashboard to understand how often you appear in the answers. 

For example, in the screenshot above, we can see that Toyota has the highest share of brand mentions on Gemini when compared to Nissan, Honda, and Mazda.

It gives you a robust picture of your influence within those responses and where you stand relative to competitors.

Citation frequency and consistency

Track how often you are mentioned in each LLM system. You might be performing better on ChatGPT than on Gemini or Perplexity. Focus your efforts on the platforms with the highest ROI and see if you can reverse engineer tactics for lower-performing answer engines.

Source mixture

This is the metric I find most revealing. If the majority of your citations come from third-party sources, it tells you that AI systems trust others rather than you. Conversely, if citations come only from your own site, you need to expand your third-party presence. 

In my opinion, the healthiest position is a balance between the two, where authoritative external sources are talking about you, and LLMs are also pulling directly from your content.

7. How should brands respond when Google controls discovery and also competes for attention?

Olabinjo Adeniran, Founder, Growth Case Studies

Google is referencing its own ecosystems, and YouTube is the clearest example, as you can see from the Moz data below.

Source: AI mode citation study by Moz

Brands need to invest in video content, whether that's working with creators, partnering with influencers, or publishing branded content on your channel. 

Video is a great way to diversify your traffic sources, and there’s evidence that it gets a ton of visibility in both LLMs and traditional search.

Charlie Clark, CEO, Minty Digital

Focus on content that produces information gain. Most content summarizes existing knowledge, which LLMs already do. AI models won’t surface regurgitated content because they can generate it without retrieval.

If you want to earn third-party mentions and disseminate content in a zero-click world, I’d advise you to invest in original research and net-new knowledge. These are two content formats that AI platforms can’t summarize or surface on their own.

8. Profound's report shows that LinkedIn is now the most-cited domain for professional queries. How can brands use LinkedIn to get more LLM citations and improve Brand Authority? 

Charlie Marchant

First, I want to clarify by saying this data only matters if LinkedIn is showing up for the prompts you care about. 

If your competitors' LinkedIn content is appearing in ChatGPT, Copilot, or whichever platforms you're tracking, go after it. If LinkedIn isn't showing up at all for your brand's queries, it's less of a priority.

For most B2B brands, LinkedIn is a key visibility channel. Founders, marketing managers, and HR leads all have voices on LinkedIn, and LLMs are citing this content.

Here’s how you can take advantage of LinkedIn for AI visibility:

Find the right voices in your company

Not everyone is active on LinkedIn, and that's fine. Start with employees who already post on LinkedIn or are interested in building a personal brand.

From what I’ve seen, corporate content doesn’t do well on LinkedIn. LinkedIn users are more likely to engage with relatable content that mirrors their lived experiences. 

Build topical pillars around your employee's job role and expertise. For example, an HR lead's natural territory is hiring, retention, staff benefits, and employee engagement, so start there.

Create content around the topics you want to own

Use a keyword research tool to identify the topics you want to rank for. Next, ask internal SMEs to publish newsletters and long-form LinkedIn posts around those themes. 

If someone is a better speaker than a writer, that's not a problem. Ask them to record a podcast, webinar, or YouTube video, then use an LLM to repurpose it for LinkedIn.

At Exposure Ninja, we've run experiments where LinkedIn posts influenced LLMs and Google AI Overviews. The two are more connected than most people realize.

9. 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

The first step is confirming whether LLMs are misrepresenting you. Large enterprise brands tend to care because they have specific adjectives and positioning statements they're protecting.

For example, a premium brand doesn't want to be described as cheap, and a specialist doesn't want to be positioned as a generalist. 

Sentiment scores are a good starting point for understanding how LLMs describe you.

Below are two examples of how this plays out in practice:

a. We ran a sentiment analysis for Beaches and Sandals, a luxury honeymoon resort, and found significant negative sentiment. 

The source was specific to grooms who arrived at their honeymoon without a tuxedo and had no option to rent one. The feedback across the web reflected their frustration, and LLMs picked it up and parroted it back. The fix was straightforward because the problem was operational, not perceptual.

b. We worked with a financial education client offering accounting and financial analyst qualifications. 

LLMs were consistently describing them as significantly more expensive than competitors, even though their pricing was identical. We recommended updating the pricing page to make the comparison clearer, and within three days, they appeared at the top of LLM responses. Nothing changed in their pricing, just how clearly they communicated.

Misrepresentation in LLMs is often a content problem, not a perception problem. Fix what the model is reading, and the output changes with it.

I’ll also encourage you to read this piece by Jamie Indigo, sharing useful tips for creating a defensive SEO strategy to protect your brand against harmful misrepresentation in AI search results.

10. If you had extra budget in 2026, where would you invest it for long-term LLM visibility?

Emma-Jane Stogdon, Organic Content Manager, Wise

Invest in offline visibility to increase Brand Authority. The brands winning in LLMs are the ones people talk about offline, and it goes beyond search optimization.

However, you can’t achieve this with siloed teams. You need an integrated marketing strategy where different teams work towards a common goal.

Your goal is to get people to talk about you in real-life conversations and recommend you to their peers.

You can do this through activation events, showcasing at industry conferences, and investing in PR. 

Adewale Adetona, Senior Product Marketing Manager, Arqiva

Invest in partnerships. We recently published white papers for two of our products in partnership with three customers, and the results exceeded anything we would have achieved on our own. 

Customer voice drove visibility, broke down barriers for our sales team, and created content that third parties wanted to reference.

Ryan Glass, Performance Experience Lead, Performics

Review your top-performing content from the last 24 months and future-proof it through a multi-channel strategy. 

Ross Simmonds has a framework called "create once, distribute forever" that applies here. 

Take content that’s working, build the information-gain version, and distribute it across formats like video, blog posts, and social. 

Repetition through distribution is how you inundate LLMs with a consistent brand message.

Charlie Marchant, CEO, Exposure Ninja

Invest heavily in digital PR and citation building to nurture brand positioning that compounds over time. 

In an article I wrote about generative engine optimization, I gave an example of The Ordinary as a brand that uses digital PR and positioning to influence perception.

Before LLMs existed, they had already built two proposition statements around best-value skincare and science-backed skincare. They published in Cosmo, Glamour, and other relevant authoritative sites, maintaining consistency for fresh coverage.

Regardless of how you prompt an LLM about affordable or science-backed skincare, The Ordinary appears in the top mentions. 

If you consistently put out the message you want, others will start repeating it for you.

Rejoice Ojiaku

Invest in training. Not deep technical training, but enough for teams to understand what the others are doing. 

The social media team should understand SEO, and the PR team should know how LLMs work. Answer engines use signals from multiple channels to generate responses. If your teams are not talking to each other, you’re leaving money on the table, and no amount of optimization will close the gap.

Remove silos within your internal marketing teams. Ensure consistent messaging across platforms and invest in a multi-channel approach to maximize visibility. 

PR, social, and SEO are not separate workstreams, and LLMs reward brands with an integrated approach.