Share article:

With the emergence of Generative AI platforms such as ChatGPT, Copilot, and Perplexity it is important to understand how they work and how your brand can appear in their generated responses. 

This presents new challenges for marketers, but it is an SEO problem.

Optimizing for Large Language Models (LLMs) is strikingly similar to traditional SEO – but with new complexities. Similar to search, appearing in LLMs requires your brand’s content to be indexable, relevant, and authoritative. Either at the time the LLM was trained or if connected to search, in real-time. 

In this article, we will jump into a quick timeline to see how LLMs and Generative AI have influenced search and how people find information. 

Then, we’ll explore why LLM optimization is an SEO problem, drawing parallels between the two disciplines and outlining strategies to ensure your brand remains found wherever people are searching. You’ll learn:

  • How AI has disrupted search
  • What is LLM optimization?
  • How to appear in Generative AI platforms
  • How to measure traffic and conversions from LLMs

AI and SEO Timeline

The launch of ChatGPT and generative AI platforms has changed SEO in a multitude of ways. As marketers are left with a rush to find the best use-cases of generative AI, SEOs are facing a new platform for Search. Today we want to focus on visibility in LLMs and what this means for the future of Search. 

There are a few key points to this timeline that are important for understanding the interplay between organic search and LLMs. 

November 2022: OpenAI Launched ChatGPT

ChatGPT shook the digital world when it was launched to the public in 2022. For search, it created a flood of AI-generated content to the internet. A month later, Google updated their guidelines from E-A-T to E-E-A-T to include (human) experience in an attempt to combat low-quality content.

While Google is not against AI-generated content, they are against it being used in spammy or unhelpful ways. 

Spring 2023: ChatGPT, Google, and Bing Launch New Products

In March 2023, OpenAI released ChatGPT-4 which significantly improved Generative AI’s output. A month later, Google launched SGE (Search Generative Experience) Beta to the US. Then in May 2023, Bing launched Copilot powered by ChatGPT-4. While these releases influenced our workflows and production capabilities, they had limited impact on organic search. 

May 2024: Google Rebrands SGE to AI Overview

By mid-2024, Google’s SGE became AI Overviews and launched in the US and 6 additional countries, with a worldwide launch planned by end of 2024. (Note, there is no confirmation of a rollout of AI Overviews to the EU at this time)

Google’s AI Overview is powered by a combination of Google’s algorithm and their LLM, Gemini. It is a generative AI response that appears in some search queries, appearing at the very top of the search results page.

This is the first major disruption to organic traffic in this timeline. Many sites are reporting significant traffic drops for some queries – especially top of funnel. Flow did a study earlier this year that examined the visibility of AI Overview for 5000 HR and workforce management keywords and also outlined strategies for ranking in AI Overviews

October 2024: ChatGPT with Search and Surpassing Bing Traffic

Just under two years from its initial launch to the public, ChatGPT surpassed Bing in traffic.

If that wasn’t enough disruption to the internet, ChatGPT with Search is released. Generative AI and Search have now combined, fundamentally changing how LLMs can act as aggregators (and competitors) of search.

YouTube

By loading the video, you agree to YouTube’s privacy policy.
Learn more

Load video

What is LLM optimization?

LLM optimization is a framework for improving the visibility of your brand and offerings in generative AI platforms like ChatGPT or Perplexity. 

If you want to appear in a user’s ChatGPT response when they query “what is the best employee management software” then you are looking for LLM optimization, also referred to as GEO (Generative Engine Optimization).

Platforms like ChatGPT are built from their own Large Language Model (LLM), a type of machine learning model designed to understand and generate human language. These models are then trained to recognize patterns and language that analyze and attempt to understand complex relationships between words and phrases. 

This process is done through the use of Natural Language Processing (NLP) which can be considered a translator that helps computers understand and respond to human language, much like how autocorrect predicts the word you’re typing.

Why LLM Optimization is an SEO Problem

SEOs have a broad skill set that position them perfectly for optimizing for LLMs. Here’s why they’re ready for the challenge:

  • Technical Expertise: SEOs have deep knowledge of indexing, crawling, schema markup, and algorithm updates, allowing them to stay ahead in AI-driven platforms.
  • Content Authority: SEOs are primed at creating content that is not only relevant but also authoritative, essential for visibility in LLM-generated responses.
  • Link Building and Digital PR: High-quality backlinks remain important for credibility, and SEOs’ skills in digital outreach can amplify a brand’s trustworthiness for both search engines and AI systems.

Tips for Optimizing for LLMs and Generative AI

Optimizing for LLMs is very similar to optimizing for search engines, but with a twist. Get started with this list of tips to help you with your LLM optimization efforts. 

Understand knowledge cut off dates of platforms

LLMs have a knowledge cut-off date. This means that the platform’s worldview, facts, and state of science or technology are limited to that date, as it was the last time the model was trained. This results in less accurate responses, with predictions and recommendations that are “frozen” in time. 

Generative AI PlatformsLLMKnowledge Cut off Date
ChatGPTChatGPT-4oOctober 2023
ChatGPT-4April 2023
GeminiGemini ProNov 2023
Perplexity AIPerplexity April 2023
ClaudeClaude 3April 2024

This is particularly important for LLMs, such as Claude, that are not connected to the internet. If your brand is struggling to appear in generative responses of these LLMs, it may take until the next update when the model has been retrained on more recent information. Similarly to how Google’s algorithm updates function. 

Some LLMs, such as ChatGPT with Search, have access to the internet, which can help supplement and inform their response. It also produces a set of linked citations and Search results (as shown below). 

Make sure your site is crawlable and indexable

This may seem overly obvious, but LLMs need to be able to crawl your website to be able to share your content. Without access to your site’s content, generative AI platforms cannot surface your brand in responses.

Verify your crawlability by checking your Robots.txt is not blocking LLMs or search engines from crawling your site. Some organizations block AI crawlers in the hopes of protecting their content from being used for training purposes. However, this might inadvertently limit visibility in LLMs. Strike a balance by carefully assessing which parts of your site are critical for indexing and which can be excluded without harming your brand’s discoverability.

Use schema markup to help LLMs understand your brand

Entities are people, places, or things defined in a knowledge graph through connected relationships. Knowledge Graphs are used by LLMs and Google in combination with other processes to understand your query in relationship to the world. 

This is done in a multitude of ways that we won’t go into, but the short of it is that we can use Structured Data Markup on our website to help LLMs better understand who we are by connecting our identities across the internet. 

For example, Google uses Organizational schema to help build its Knowledge Graph. Including this markup on your homepage can help not only Google, but also LLMs in connecting your brand identity across the internet. 

The image below shows ClickUp’s Knowledge Panel which signals that the brand is part of Google’s Knowledge Graph. 

Schema.org’s Markup Validator can show you how some sites are using Organization schema to connect their identity across the web. 

Get mentioned on high value, relevant lists and domains

Link building has made a bit of a comeback and a rebrand in 2024 as Digital PR. Specifically for SaaS brands, LLMs use listicles as a way to quickly qualify and aggregate best of software lists and recommend solutions. It is more important than ever to be mentioned on multiple of these high value domain lists and review directories. 

Additionally, it is important that we align brand messaging and imagery across mentions. It is possible that an incorrect mention on a high-value domain can be used in an LLMs response. To avoid this from happening, it is important to review and update brand mentions whenever necessary. 

Be careful not to put all of your eggs in one basket. Forbes Advisor was previously a top source for LLM listicles (due to their high rankings in Google) but as of November 2024, many of its pages have been removed from the search index and tens of millions of organic sessions have dropped to under 1 million.

Align brand messaging across web properties

Along with aggregating data, LLMs are also able to assess sentiment of written content. Meaning that if reviews of your product are negative or outdated on platforms like Reddit, an LLM may describe your brand unfavorably.

This can particularly become a problem when potential customers are searching ‘[your brand] pricing’ or ‘[your brand] reviews’ and land on a Reddit forum full of outdated information. 

Conversely, glowing reviews on sites like Getapp, Trustpilot, or other industry relevant sites can positively shape AI responses.

Our Branded Search Checklist template can help you audit the current state of your brand in Search to help align brand messaging across web properties. 

We recommend engaging in forums where user generated content is being shared, commenting on socials, and even creating pages or blog posts on your site that talk about your brand. It can be as simple as the top 10 FAQs potential customers are asking the sales team in the past 90 days.

If you want to take branding into your own hands and even go a step further, you can offer your insights as a trusted Subject Matter Expert (SME). There are numerous strategies for this, but the most proven success stories come from responding to queries from journalists who are hoping to include information that will make their content more interactive and have better depth. It also creates a sense of community.

Our List of HARO Alternatives can help you navigate the different platforms for connecting sources to writers. Just remember that when you supply a quote, you are representing, and in some cases, creating your brand. Once you have a successful strategy in place, you’re likely to gain placements on high-authority sites, thereby establishing your company as an expert in its field.

Rank well organically across Search engines

Ranking well in organic search engines positively influences visibility for LLMs when connected to Search engines. For ChatGPT and CoPilot that is Bing, while Perplexity and Gemini use Google. 

Generative AI PlatformsInforming Search Results
ChatGPTBing
GeminiGoogle
PerplexityGoogle
ClaudeNot connected to internet

This results in diversity in responses across platforms. Below are examples of using ChatGPT and Perplexity for the same exact query “fun ways to welcome new employees” which produced fairly similar results, but different citations. 

ChatGPT with Search is informed with Bing search results. The generative response is a list of ideas with links to citations. This particular query had 5 citations, which can be seen on the right side panel. Below citations, there is a continuation of links as part of ChatGPTs “search” functionality. The majority of these links presented are ranking on page 1 and 2 of Bing search results for the exact same query. 

Repeating the process inside Perplexity, which is connected to Google search, we will see a very similar output including sources and a list of ideas. The sources shown and used to inform the generative response are all from page 1 of Google search results for the same query. 

LLMs like ChatGPT and Perplexity are fundamentally tied to search engines. As a result, they aggregate and summarize content from top-ranking search results to generate responses. 

This close relationship between LLMs and search highlights a shift in how users discover information, brands, and websites. Instead of directly visiting search engine result pages (SERPs), users may increasingly rely on AI-driven summaries that reshape the path to visibility.

Tools and Resources for LLM Visibility and Traffic Insights

Unfortunately we are unable to track mentions or see impressions from LLMs, but we do have a few tools at our disposal to help gain insights into visibility and even traffic.  

HubSpot’s AI Search Grader

Hubspot launched a free AI Search Grader that allows you to enter your brand, location, and category to generate a report for brand share of voice and sentiment in OpenAI and Perplexity. This is a great report to start with when strategizing for LLMs to understand how competitors in your category are faring in AI. 

Track Referral Traffic in GA4

With the launch of ChatGPT 2 years ago, there were no links in its generated responses. Today with ChatGPT with Search there are two sets of links; citations and search results. 

When a user clicks on a link and lands on our website that is referral traffic. We can track that through GA4 by creating an Explore Free Form report using a filter for Page Referrer to contain this regex: 

^https:\/\/(www\.meta\.ai|www\.perplexity\.ai|chat\.openai\.com|claude\.ai|chat\.mistral\.ai|gemini\.google\.com|bard\.google\.com|chatgpt\.com|copilot\.microsoft\.com|edgeservices\.bing\.com)(\/.*)?$

LLM Looker Studio Dashboard

If you are anything like us at Flow, we love a good Looker Studio dashboard and so we made one to help you check and monitor traffic to your site from LLMs. The LLM Dashboard highlights which platforms are driving the most traffic and conversions and which pages users are landing on. 

LLMs as part of an integrated marketing strategy in 2025

AI is becoming more embedded in how people discover and interact with content online, and this shift is set to accelerate in 2025. LLMs like ChatGPT, Perplexity, and Gemini rely heavily on search engines to inform their outputs, which means your search rankings now play a dual role: driving direct traffic and influencing the narratives shaped by generative AI.

For marketers, this changes the game. Fresh content will remain essential, not just for ranking but for staying relevant in the AI-driven summaries users rely on. Diversifying traffic sources should also be a priority to avoid over-reliance on any single channel, especially as LLMs reshape referral pathways. On top of that, regularly monitoring how your brand appears in AI outputs will help you catch missed opportunities or inaccuracies before they scale.

This is where SEO shines. With expertise in making content discoverable and adaptable to evolving search behaviors, SEOs are uniquely positioned to bridge the gap between traditional search and AI. By taking a proactive, integrated approach, marketers can ensure their brands stay visible and competitive in a future where search and AI are inseparable.

Get in touch!

Send us a message to schedule a call to chat about your LLM traffic.

Learn more

Author

Helene Jelenc
Helene is one of our SEO specialists with a background in anthropology. When she isn’t writing about the interplay of culture and search, you can find her traveling to old cities and hanging out with her chickens.
Flow Blog

You may also like: