Before Google dominated our digital lives, my search engine of choice was ‘Ask Jeeves’ (I know I’m aging myself!). I remember the exact moment I searched ‘what is a blog’ into Ask Jeeves, because that one search so many years ago set the trajectory to where I am today. When Google surpassed early search engines like Ask Jeeves and Dog Pile, I spent a lot of time testing and poking and prodding to see how I could get the content I was writing to appear high up in search results. As it turns out, what I had originally thought of as “gaming” Google became a massive industry known today as search engine optimization or SEO for short.
Now we are seeing the next manifestation of online information retrieval in the form of AI Large Language Models (LLM’s) like Chat GPT and Google’s Gemini, and the generative search engine Perplexity. Over the past year, ChatGPT has exploded onto the market, reaching over 300 million weekly active users as of December 2024, which is approximately 1.2 billion monthly active users.
Google’s Gemini, rebranded from Bard in February 2024, reported 330.9 million monthly users as of January 2024.
While Perplexity has much smaller numbers in 2024 (15 million active users on its website and app), I suspect user growth will explode in 2025 as more people find out about it and use Perplexity as a substitute for Google. As you might guess, I’m a really big fan of Perplexity and highly recommend trying it out.
These numbers are impressive and have momentum as these LLM’s become more user-friendly and integrated into organizational workflows. But they pale in comparison (for now at least) compared to Google which processed approximately 8.5 billion searches daily, totaling around 3.1 trillion searches in 2024.
So for now at least, Google remains on top when it comes to people seeking information online, and search engine optimization will continue to be a viable means of ranking high as an online source of information. Of course, Google is well aware that there is a shift in the way people are finding information online and they are not just sitting around watching while people move from Google search to LLM queries.
At the 2024 New York Times DealBook Summit, Google CEO Sundar Pichai addressed the future of search and Google saying that:
“Search itself will continue to change profoundly in [2025]. I think we are going to be able to tackle more complex questions than ever before. I think you’ll be surprised even early in [2025] the kind of newer things search can do compared to where it is today.”
And,
“If you look at the last couple of years, we have, with AI overviews, Gemini is being used by over a billion users in search alone. I just feel like we are getting started.”
Clearly Google is not sitting on their heels and I have no doubt we will continue to see deeper integration of their LLM Gemini into Google search functionality, as well as their other major products like Google Workspace. I think long term, Google has a very good chance of continuing to be a market leader in online search if they play their cards right. By integrating Gemini into Google search as well as into their other products like Workspace, their built-in user base is so massive that it will be hard for companies like Open AI (Chat GPT) to keep up. Google is a verb after all.
Given all this, we are seeing the early dawn of a new digital marketing practice that some are calling Answer Engine Optimization, while others (including me) are calling Generative Engine Optimization (GEO).
What is Generative Engine Optimization?
The term Generative Engine Optimization was coined in a June 2024 paper titled: GEO: Generative Engine Optimization. Here is the original PDF version: GEO: Generative Engine Optimization, June 2024.
Developed by researchers from Princeton University and the Indian Institute of Technology Delhi, the GEO framework outlined in this paper provides key insights into how LLM’s like Claude, Chat GPT and Perplexity are using online sources in user queries and provides advice on how content creators and digital marketers can produce their content in a way that increases their chances of appearing in LLM search results.
Early adoption of GEO best practices offers a huge opportunity for those of us working in the campaign and advocacy space. Establishing your organization as a credible source in AI systems helps amplify your message and as more users rely on AI for information discovery, GEO becomes a path to connect with those who might never visit your website directly.
Luckily, for those already somewhat fluent in SEO best practices, the jump to GEO is not that difficult. In my experience so far with GEO, the suggested best practices are actually improving the content I am publishing. If I was to summarize GEO in a single sentence it would be:
Create well-written and structured content, with clear evidence for your claims including (where appropriate) statistics, data citations and/or first hand quotations or accounts by real people that add a narrative context to the information you are providing.
So what can we do to get our content included in Chat GPT and other LLM search queries?
The 2024 GEO research tested numerous optimization strategies and identified several approaches that significantly boost content visibility in generative engine responses.
Much like SEO, it is important to remember that we are still (for now) writing content for real-live people and not just for Google and Chat GPT. So when it comes to implementing the methods suggested below, find a balance between writing for people and writing for the bots. Do not sacrifice your unique writing style and voice for the bots if it comes at the expense of writing for real people.
What follows are concrete suggestions provided by the GEO paper authors on how we can produce content that is AI-friendly. The authors claim that these methods have the potential to boost your content’s visibility in AI-enabled search by up to 40%:
Statistics integration: Generative engines appear to prioritize content with concrete data points over qualitative discussion.
For those writing on social and environmental issues, this means replacing general statements with specific, verifiable statistics whenever possible. Instead of saying “many people are affected by climate change,” specify “over 20 million people are displaced annually by climate-related disasters according to the UNHCR.” Another example, rather than stating “Homelessness is a serious problem in our city that affects vulnerable populations,” write “Homelessness affects 12,500 residents in our city, with 68% belonging to historically marginalized communities according to the 2023 Housing Coalition census.”
Quotation addition: the GEO authors claim that including quotations demonstrated the highest performance gains of all methods tested, increasing visibility by over 40% across various queries.
For those in the campaign and advocacy space, consider incorporating direct quotes from affected communities, experts, or respected figures in the movement to add authenticity and authority to content. For instance, instead of simply stating “Our organization believes clean water access is a fundamental human right,” strengthen this by adding “Our organization believes clean water access is a fundamental human right. As UN Special Rapporteur on the human rights to safe drinking water and sanitation Pedro Arrojo-Agudo states, ‘Water is not a privilege, but a basic human right essential for human dignity and survival.'”
Citation integration: by including citations, the researchers saw substantially improved content visibility, reinforcing the value of demonstrating research thoroughness and accuracy, with improvements exceeding 30%.
Content should try to cite respected peer-reviewed research, reports from reputable organizations, government data, or other credible sources to back up claims. Multiple citations from diverse sources appear to perform better than single citations. For example, a statement like “Racial disparities in healthcare outcomes persist despite policy efforts” could be edited to say, “Racial disparities in healthcare outcomes persist despite policy efforts, with Black Americans experiencing 2.3 times higher maternal mortality rates compared to white Americans (CDC, 2023) and receiving fewer pain management interventions for equivalent conditions (Journal of Pain Research, 2022).”
Fluency optimization: resulted in significant visibility gains of 15-30%, indicating that generative engines value not just informational content but also its presentation quality. Ensure your content is well-structured, grammatically correct, and flows logically. Consider having professional editors review your key content pages.
An especially valuable insight from the GEO research is that optimization strategies vary significantly in effectiveness across different domains and query types. For debate and opinion content, authoritative tone modifications and statistics additions showed the strongest results. Environmental and scientific content benefited more from fluency optimization and technical terminology. Legal and policy content saw the greatest improvements with citation incorporation and statistics addition. People-focused and human interest content responded best to quotation additions.
So how can I get started with Generative Engine Optimization?
To implement GEO effectively for your content, start by auditing your existing content to identify your most important pages that address high-stakes issues where your voice needs to be heard. You can use Google Analytics for this, by going into and looking at the pages your site receives the most traffic to over the past few years. I would look at this both from a quantity and quality perspective as their will be pages on your site that might not get a large volume of traffic, but instead attract more valuable visitors, so be sure to look at other metrics like conversions, time-on-page and page-per-visit metrics.
Prioritize optimization efforts by beginning with your most essential content that addresses frequently searched topics. Apply domain-specific optimizations based on your content type, for instance: policy papers should focus on citation integration and statistics, personal stories should emphasize quotations and fluency, position statements should strengthen authoritative tone and citation and scientific explanations should improve technical terminology and fluency.
How do I know if my GEO efforts are successful?
For now, it is a little tricky when it comes to knowing exactly how many times AI-tools like Chat GPT, Gemini and Perplexity are directly querying your website and content as a source. Right now, you can look in your traffic logs and source reports to see if anything is labeled in a way that identifies AI activity. For instance, look for identifiable bot traffic and Internet Service Provider (ISP) labels. We know that Chat GPT relies on Microsoft Azure servers in San Antonio, Texas, so a surge in traffic from those data centers could be an indicator.
I suspect that as GEO grows as a digital marketing practice, we will see better access to software that will allow us to easily and more accurately record traffic from AI tools.
In the meantime, I would suggest firing up a few of the major LLM’s, like Chat GPT, Claude, Gemini and the generative search engine Perplexity and start running queries to see if these tools are including your content as a source. Models like Perplexity provide real-time processing information so you can see what sources it is considering, and its thinking process for how it got to its final answer. Look to see the questions Perplexity is asking itself on a given topic and see if you might have content that could be optimized for those questions. If you don’t, and it is something you and your organization are expert in, consider writing a GEO-optimized page to answer that question.
For advocacy organizations and social justice content creators, GEO represents more than just a technical adjustment. This is a huge opportunity to get out front and ensure that your important perspectives are not algorithmically marginalized in the AI era. The rise of generative engines threatens to further concentrate power in the hands of large tech companies and mainstream sources. By implementing effective GEO strategies, advocates and campaigners can help democratize the digital information landscape and ensure diverse voices are represented in AI-generated responses.
As we navigate this paradigm shift, investing time in optimizing your content for generative engines is not just about maintaining visibility, it’s about ensuring that your community’s voice, perspective, and lived experiences remain accessible in an increasingly AI-mediated information ecosystem.
Sources:
The Verge: ChatGPT now has over 300 million weekly users
Softonic: Google Gemini Stats 2024: How Many People Use It And Key Traffic Insights
Business of Apps: Perplexity Revenue and Usage Statistics (2025)
Blogger Passion: 74 Google Search Statistics In 2024
Search Engine Land: Google CEO Sundar Pichai: Search will profoundly change in 2025
Agio: Where is ChatGPT Hosted?