Get AI to Talk About Your Brand

Large language models (LLMs) are AI systems trained on vast amounts of written text to understand and generate human language. 

LLMs power generative search and conversational search:

  • Generative search: A form of search that generates a natural-language response to the user’s query (e.g., Google AI Overviews)
  • Conversational search: A form of generative search that allows back-and-forth dialog between the user and the LLM-based tool (e.g., ChatGPT, Google AI Mode)

Many people refer to LLM-powered search tools simply as LLMs. So I’ll do the same.

In this guide, you’ll learn why and how to optimize for LLMs.

What Is LLM Optimization?

LLM optimization (LLMO) is a marketing tactic that aims to improve a brand’s visibility and portrayal in LLM-generated responses—like those found in ChatGPT, Google’s AI Overviews, and Google’s AI Mode.

Key LLM optimization techniques include:

  • Getting positive mentions of your brand on reputable websites (so LLMs understand your brand better and perceive it more favorably)
  • Creating original, useful, and LLM-friendly content (so LLMs are more likely to cite your content in their responses)
ChatGPT conversation is labeled with the user prompt, LLM response, brand mention, and citations.

Most of the tactics used in traditional search engine optimization (SEO) are good for LLMO, too. 

But you can use other techniques to improve your LLM visibility. More about those later.

Why Does LLM Optimization Matter?

LLM optimization matters because it allows you to improve your brand’s visibility and portrayal in LLM-generated responses—content that people increasingly see and engage with.

Plus, the average AI search visitor is 4.4x more valuable than the traditional organic search visitor, according to Semrush’s AI search study

LLMO can ultimately help you:

  • Increase brand awareness
  • Protect and improve your brand’s reputation
  • Generate more revenue

LLMO could give your brand an immediate boost. But the long-term benefits may be more significant.

After all, AI adoption is set to increase. And information about your brand can become embedded in the training data used for future and new versions of LLMs.

This means the work you do now could ensure your brand is portrayed more frequently and favorably in LLM responses in the future.

How Frequently Do Users See LLM Responses?

Google users in the U.S. saw LLM responses on 13.14% of search results pages in March 2025, according to Semrush’s AI Overviews Study.

These LLM-generated responses are AI Overviews, and they often appear right at the top of the page. Like this:

AI Overview is denoted by a diamond shape icon and include citations on the right-hand side.

Many users seek out LLM-generated responses by engaging with conversational AI tools like ChatGPT, Microsoft Copilot, Perplexity, and Claude. Which collectively attracted over 600 million unique visitors in May 2025, according to data from Semrush’s Traffic Analytics tool.

Unique visitors metric is highlighted in Traffic Analytics for mentioned AI tools.

Keep in mind that users’ exposure to LLM-generated responses is set to grow considerably as AI adoption and investment increases. 

For example, Google recently launched AI Mode, which provides a ChatGPT-like experience within the Google search interface.

LLMs are set to become a major revenue and traffic driver by 2027, according to Semrush’s AI search study.

8 LLM Optimization Techniques

No one fully understands how LLMs decide what brands to mention or cite in their responses, but based on our current understanding of how LLMs work, we believe these LLM techniques could improve your brand’s visibility.

1. Get Brand Mentions on Commonly Cited Websites

Identify sites or pages in your niche that LLMs commonly cite. Then work to get your brand mentioned on those sites or pages.

That way, LLMs may be more likely to include your brand in their responses.

In the example below, ChatGPT cites product recommendations from a TechRadar blog post. 

A brand citation in ChatGPT is pulled from a brand mention.

Quora and Reddit are the most commonly cited websites in Google AI Overviews, according to Semrush’s AI search study.

So, engaging with these platforms and mentioning your brand where appropriate could play an important role in LLM search optimization.

The most cited domains in Google AI Overviews are predominately Quora and Reddit then LinkedIn, YouTube, New York Times, and more.

To identify commonly cited websites in your niche, perform relevant searches on AI platforms and keep a record of the citations you see. 

Or use a specialized tool like Semrush Enterprise AIO. Which can identify the most mentioned sources for relevant prompts.

Improve AI Visibility Summary shows brand portrayal across sources as well as improvement potential.

You can use various digital PR and link building tactics to get your brand featured.

Chris Tweten, Chief Marketing Officer at Spacebar Collective, uses this tactic to secure valuable exposure for his clients.

He says: 

“We’re measuring success by tracking which pieces of media we produce later become citations in ChatGPT, Gemini, and Perplexity. From here, we’re also tracking total traffic coming from those platforms as well as actual conversions off of that traffic.

“Traffic from ChatGPT is converting at around 30% for one of our clients.”

2. Establish a Unique Value Proposition

Establishing a unique value proposition (UVP) that sets your business apart from competitors could give LLMs more reason to mention your business.

This LLM optimization tactic could be especially important if you’re a retailer or operating in a saturated market.

Say a user asks where to buy a relevant product type.

If you have a strong UVP (e.g., a better product range or lower prices), LLM-powered models may be more likely to suggest your business.

Like this:

ChatGPT responds with their recommended retailers, including why to buy from that brand.

For this tactic to work, clearly and consistently communicate your UVP in spaces LLMs could learn from. Such as your website, online communities, industry websites, and social media.

3. Create Original and Useful Content

Create original and useful content for LLMs to cite when relevant users submit relevant prompts. 

Originality may give the LLM a reason to mention or cite you instead of someone else. While usefulness ensures your content meets relevance and quality standards.

I recommend creating content tailored to highly specific user needs or intents. Because users often submit complex prompts in conversational AI—and competitors may not be addressing these needs yet.

Also try to cover all stages of the marketing funnel:

Top of the funnel: people become aware of the problem you can solve. Middle: Prospects wants a solution and consider their options. Bottom: Prospects decide on a solution and become customers.

Top-of-the-funnel audiences are likely to use conversational search because it can provide impartial and personalized guidance.

These users may then continue their learning journeys within the chat interface. Meaning they don’t visit your website or interact with other marketing channels until they’re ready to convert.

The more the LLM mentions or recommends your brand during this journey, the more customers you’re likely to attract.

4. Make Your Content Easily Accessible to LLMs

To ingest and cite your content, LLMs must be able to reach and process your content.

Here’s how you can make content more technically accessible to LLMs: 

  • Prioritize server-side rendering. LLMs generally learn from the raw HTML of a webpage—not content that appears after JavaScript runs. So, minimize reliance on JavaScript rendering.
  • Ensure public accessibility. LLMs can only train on publicly accessible content—they can’t access content that’s behind a paywall, login wall, or AI-restrictive license
  • Follow technical SEO best practices. LLMs sometimes retrieve information from search engine databases, so make sure to follow technical SEO best practices

Jeremy Howard, Co-Founder of Fast.ai, has proposed that the industry adopt a standardized file to provide information to LLMs, called llms.txt. (This would work similarly to robots.txt, a file designed to provide information to search engine crawlers.)

However, AI models have not formally adopted the file format yet.

5. Incorporate Image and Video Content

Incorporating images and videos into text-based content could improve your LLM-based visibility for a few reasons.

First, multimodal content tends to be more information- and context-rich, which can improve machine understanding. And reinforce your expertise.

Second, many LLMs support multimodal searches.

Google AI Mode, for example, lets users upload images and ask questions about them.

Image and "what is this called?" are entered as the prompt. Google AI responds with citations and says the image is a double rainbow.

LLMs may be more likely to retrieve and cite content that includes similar imagery to that used in the prompt. As long as the image is accompanied by explanatory text. 

Third, many LLMs cite various content formats in their responses. 

For example, Google’s AI Overviews sometimes include linked image results:

AI Overview for a question query shows an image in the result which links to the article it came from.

6. Optimize Content at the Passage Level

Optimizing content at the passage level could be helpful because LLMs often use passage-level retrieval (meaning they look for the most relevant segments of text rather than the most relevant documents).

Here are some suggested techniques for optimizing content at the passage level:

  • Be specific. Use clear and precise wording throughout your document, making sure to mention relevant entities. This can help LLMs and users extract more meaning from your content.
  • Avoid external dependencies. Ensure key passages and sentences make sense in isolation. Avoid dependencies on earlier passages or external content.
  • Stay on topic. Focus each passage around a specific topic or subtopic. Avoid asides, which can confuse LLMs and distract users.
  • Follow a logical structure. Introduce ideas and concepts in a logical order. And use subheadings to group closely related passages. 

Passage-level retrieval may also make document depth less important. Because the LLM can extract content from multiple documents on your site.

This means it could be better to create separate documents tailored to highly specific use cases, rather than comprehensive documents covering multiple use cases.

7. Tailor Content to Target Audiences

Tailoring content to a specific target audience could be important in LLM search optimization because many LLMs can provide highly personalized responses based on past interactions with the user. 

The better your brand or content aligns with the user’s profile and intent, the more likely you may be to get mentioned or cited in the LLM response.

For example, if a user has memory enabled, ChatGPT may learn over time that the user is a student.

When this user later asks for product recommendations, ChatGPT may be more likely to mention and cite brands that explicitly target students in their marketing efforts. Like this:

Landing page dedicated to Canva for students.

You can learn more about your target audience with the Semrush Traffic & Market Toolkit.

8. Manage Your Online Reputation

Online reputation management could be an important aspect of LLM optimization because LLMs build an understanding of your brand based on how it’s portrayed online.

If customers or influencers talk negatively about your brand, you may see these views reflected in LLM responses. 

Equally, positive ideas can be repeated to your audience.

Here are some ways to improve your brand’s online reputation:

  • Respond to reviews quickly and strategically
  • Encourage happy customers to leave online reviews
  • Invest in providing great customer service
  • Collect and act upon customer feedback
  • Prepare a crisis management strategy
  • Use digital PR and influencer marketing to secure positive brand coverage
  • Work to remove inaccurate or unflattering content about your brand

With Semrush Enterprise AIO, you can learn how popular LLMs portray your brand’s strengths and weaknesses. 

Brand Strength and Weaknesses report shows breakdown with categories like "robust analysis," "high pricing," "unique tools," and more across LLMs.

You can then work to amplify your strengths and mitigate your weaknesses.

Start Tracking LLM Search Visibility

To track your LLM search visibility effectively, measure how frequently and prominently your brand appears for relevant personas using relevant prompts across various models.

  • Relevant personas: AI responses can be highly tailored to the user, so it’s important to mimic and measure visibility for each audience segment
  • Relevant prompts: AI prompts tend to be far more specific and conversational than search engine queries, so it’s not possible to find and track keywords like you can in traditional SEO. Instead, choose prompts to represent conversations in key topical areas.
  • Various models: The AI search market is less concentrated than the traditional search engine market—at least for now. You can benefit from tracking your visibility across multiple models.

Also measure brand sentiment and positioning in LLMs to ensure alignment with your goals. As negative or inaccurate portrayals can do more harm than good.

Semrush Enterprise AIO generates brand-specific prompts to cover your entire buying journey. The tool then uses these prompts to measure your visibility and portrayal across LLMs.

AI Performance Overview shows share of voice and brand visibility in AI tools over time as well as compared to competitors.

The tool also provides competitor insights and details the most commonly cited pages.

If an enterprise tool isn’t the right fit for your LLM optimization strategy, use the Semrush AI Toolkit instead. 

Leave a Reply

Your email address will not be published. Required fields are marked *