In May 2026, Meltwater relaunched its website. Within weeks, AI citations across 50 tracked pages increased by 73%, rising from 99,675 to 172,862.
The results offer a practical look at AI citation optimization. This includes the process of making website content easier for AI search engines and large language models to discover, understand, and reference as a source.
Much of the existing advice on AI visibility remains broad or theoretical. This article breaks down what Meltwater changed, which pages generated the strongest results, and what the data reveals about improving citations across different AI platforms.
Contents
Why AI citation optimization is becoming a core discipline
Inside the case study: how a website relaunch drove a 73% citation jump
The content signals that actually drive AI citations
How to apply this to your own site
Track and repeat: make citation growth measurable
FAQs about AI citation optimization
Why AI citation optimization is becoming a core discipline
Search visibility is no longer determined only by where a page ranks. Now an AI assistant and Google AI Overviews read across the web and return a synthesized answer, often with a handful of sources cited inline. That's the shift behind generative engine optimization, and it changes what you're optimizing for.
The impact on search behavior is already measurable. Pew Research Center analyzed 68,879 Google searches and found that users clicked a traditional search result on just 8% of pages with an AI summary, compared with 15% of pages without one. Only 1% clicked a source cited within the AI summary itself.
That system is getting more traffic every year. Traffic to AI chatbots grew nearly 81% year over year, which means the pages that win citations are increasingly shaping what users discover before they ever reach a traditional search results page. That makes citation optimization a necessary part of broader AI search optimization and a complement to traditional SEO.
Inside the case study: how a website relaunch drove a 73% citation jump
Meltwater relaunched its core marketing site in May 2026, and the timing wasn't only about a design refresh. According to VP of Global Marketing Jenny Force, the company had already spent eight years building disciplined SEO investment, putting it in the #1 LLM visibility position among its competitive set before the relaunch even started. The question behind the relaunch wasn't how to get noticed. It was how to extend a lead competitors were actively trying to close.
That goal shaped how the team ran the project. TJ Kiely, director of brand strategy and integrated marketing, led the messaging and positioning. Saskia Sarışın, head of SEO/GEO, brought expertise in how LLMs interpret and cite content.
Sarah Mai Stokholm, global head of digital, co-led the GEO/AEO strategy across site navigation and conversion paths. Chris Hanson, director of brand experience, translated that strategy into the live site. As Force put it, “The window to build a real lead is narrowing.”
After the launch, the team analyzed LLM visibility across 50 key pages, comparing May 1 through June 4 with the previous period:
- Total citations grew 73%, from 99,675 to 172,862 across the 50 tracked pages, a gain of 73,187 net new citations in a matter of weeks.
- New pages drove 86% of that growth. Pages built around Meltwater's capabilities, platform, and use cases made up just 23 of the 50 analyzed URLs but added more than 62,000 net new citations between them.
- The homepage gained nearly 13,000 citations, moving from 9,040 to 21,931, which improved its LLM ranking position from 57 to 15.
- Individual capability pages saw substantial lifts: Social Listening gained 11,691 citations, Platform gained close to 10,000, and Media Intelligence, Measurement & Analytics, Consumer Intelligence, and Influencer Marketing all posted meaningful growth. Some existing pages, including Media Monitoring, Influencer Management, PR Reporting, and Consumer Insights, gained too.
- A handful of pages declined, a reminder that this kind of work isn't a one-time project.
The broader pattern matters as much as the individual results. LLMs are more likely to surface pages that clearly explain a topic or capability in depth. The new capability pages gave them clearer information about what Meltwater offers across different stages of the buyer journey.
The content signals that actually drive AI citations
Six patterns show up consistently in the pages that gained the most citation volume.
- Write self-contained sections: Each section should make sense on its own. AI systems may pull a single paragraph or passage from a longer page, so sections that depend heavily on earlier context are harder to use in generated answers.
- Lead every section with the direct answer: Start with the answer, then add context. A section titled “What is share of voice?” should open with a clear one- or two-sentence definition before moving into details or caveats. This makes the core information easier to identify and cite.
- Back claims with specific stats and named sources: “Significant growth” gives an AI system little concrete information to reference. “A 73% increase in LLM citations” provides a precise figure and, when linked to the original source, evidence you can verify. Use specific numbers where they add value, name the source, and link to the original research.
- Use consistent terminology across pages: Use the same terminology across every page in a topic cluster. This helps AI systems connect brand mentions to the right products and entities, strengthening brand authority, entity authority, and topical authority. It also makes LLM metrics and KPIs easier to compare across the site.
- Add visible FAQ content, not only FAQ schema: FAQ schema can help machines understand a page’s structure, but it does not replace visible question-and-answer content. The written answers on the page provide the information an AI system can quote, summarize, or cite.
- Know which engines you're optimizing for: ChatGPT, Claude, Gemini, and Perplexity do not source or weigh content the same way, and their citation behavior changes with model updates. Treating AI citations as a single metric can hide where visibility is growing or declining, as Meltwater’s page-level data showed.
How to apply this to your own site
You don't need a full relaunch to start seeing citation gains. The same signals that drove Meltwater's results apply page by page, and they hold up across the broader GEO strategies teams are testing right now.
Audit your highest-traffic, highest-intent pages first
Start with the pages that already bring in visitors or help buyers evaluate your product, such as product, use-case, comparison, and pricing pages. These pages give you the best chance of seeing whether your changes affect AI citations because they already have search visibility and clear commercial value.
Review each page against the questions a buyer would actually ask. Does it clearly explain what the product does, who it is for, how it works, and what outcome it delivers? Add direct answers where they are missing, break dense sections into self-contained blocks, and support important claims with evidence.
Before publishing the changes, record how often AI platforms cite each page, and which AI platforms cite it. Then compare the same pages after the updates to see what improved. Also check your robots.txt file to make sure you're not unintentionally blocking AI crawlers such as GPTBot and PerplexityBot.
Map the sub-questions AI tools ask around your core topics
Go beyond the main keyword and identify the specific questions someone might ask an AI assistant. For a product that solves a workflow problem, this could include how the process works, common challenges, use cases, comparisons, and expected outcomes. Make sure the page answers each question directly rather than forcing the reader or the AI system to piece together the answer.
Rewrite brand-heavy pages into capability-first pages
Brand-heavy pages often focus on broad claims such as innovation, leadership, or all-in-one solutions. That language may sound polished, but it gives AI systems little specific information to cite.
Rewrite these pages around the capability itself. Explain what it does, who uses it, how it works, and what result it helps deliver. Include concrete use cases, workflows, and examples instead of relying on general brand statements.
Meltwater’s results show why this matters. Its new capability, platform, and use-case pages drove 86% of the total citation growth after the relaunch.
Get your SEO, digital, and brand teams working from one plan
More than copy alone shapes AI citation performance. Messaging, page structure, navigation, internal links, and technical setup all affect how easily an AI system can understand and use a page.
That is why SEO, digital, and brand teams need to plan the work together from the start. Meltwater’s relaunch combined search and GEO expertise with brand strategy, site architecture, and design, rather than treating them as separate stages. This helped ensure that each page was clear, consistent, easy to navigate, and built around the questions buyers actually ask.
Track and repeat: make citation growth measurable
There's a real gap between hoping AI systems cite your content and structuring it so they will, and between structuring for citations and knowing whether it's working. Citation behavior varies by platform, and it can vary by page even on the same site, so tracking your LLM visibility needs to happen at a granular level, not as one blended number.
This is what Meltwater's GenAI Lens is built for. Since Meltwater introduced GenAI Lens, teams have used it to see which AI systems cite which pages, how rankings shift over time, and where the biggest gaps sit. It's the tool Force's own team used to run the analysis behind these results, and to catch the handful of pages that lost ground even as the site overall gained. How you report on LLM visibility internally matters just as much as tracking it, since this kind of work needs ongoing buy-in, not a one-time readout.
FAQs about AI citation optimization
What is AI citation optimization?
AI citation optimization is the practice of structuring website content so large language models can find, understand, and cite it as a source in AI-generated answers. It focuses on direct answers, self-contained sections, and specific sourced data rather than ranking position alone.
Is this different from traditional SEO?
Yes, though the two overlap and often sit under the same broader answer engine optimization strategy. Traditional SEO targets ranking position on a search results page. AI citation optimization targets whether a model chooses to cite your page as evidence for its answer, which depends more on how clearly and directly your content answers a specific question than on backlinks or keyword density alone.
Does schema markup guarantee more citations?
No. Schema markup helps machines parse your page's structure, but it doesn't create citable content on its own. The visible text, especially direct answers and specific FAQ content, is what LLMs actually pull from. Schema supports the content; it doesn't replace it.
How long does it take to see citation growth after structural changes?
In Meltwater's own case, citation volume grew significantly within about five weeks of the relaunch going live (May 1 through June 4). Timelines vary by how frequently a given LLM re-crawls and re-indexes your content, so tracking citation data by page and by platform over several weeks gives a clearer read than checking once.
