SEO vs GEO: Are These Really the Same Strategy?
SEO for AI (GEO/AEO/etc.): How is it different from the traditional SEO?
Generative Engine Optimization (GEO) is a new term suggested by many experts to refer to tactics aiming at increased brand visibility in AI Answers (as well as traffic from those, which, in my opinion, is a lower priority). There are a few alternative terms floating around, including AEO, AIO, LLMO (LLM Optimization), etc. - all referring to AI visibility optimization.
There’s an ongoing debate about whether we need new terms for that because, in lots of aspects, search and AI optimization tactics overlap.
👉 JOIN US to discuss this topic LIVE on LinkedIn
👉 Join my Subreddit covering SEO for AI trends, tools, tests, etc. We are breaking Reddit by building a friendly (and helpful) subreddit :)
Whatever it is called in the future, AI / LMM optimization is not quite the same as SEO (even though SEO is fundamental), and that comes from the core differences between search engines and generative AI engines:
LLMs may not search to find answers (they surely search more these days than they did a few months ago). They would often generate answers based on their existing training data. The training data doesn’t store URLs. It is simply a structured knowledge base, just like a human mind that can recall answers but not necessarily the sources.
AI platforms don’t index or cache URLs. So when LLMs have to search, they rely on third-party search engines (Google, Bing, Reddit, etc.).
When searching, AI agents “fan out” outside of the initial prompt. When fanning out, they will refer back to their training data to decide how to do it (so it's being in training data again!)
After the search is performed, to generate an answer, AI agents “visit” each web page to “read” it and fetch the answer from its content. AI crawlers are not as advanced as search crawlers, so they won’t be able to render a web page as fast or easily as Google would.
The nature of prompting is different than querying a search engine. People tend to prompt in “tasks”, e.g., “find that”, “compare”, etc. Sooner or later, that will be “buy.” It is not quite searching for answers.
To make it simpler, let’s take a look at a local query and how it will be treated by search and AI:
Let’s say someone is looking for fish tacos in an area…
Google search: Pulls map results of all businesses mentioning “fish tacos” in the description, on the site, or in the reviews.
ChatGPT/Gemini: Pulls map results of all businesses mentioning “fish tacos” in the description, on the site, or in the reviews. BUT ALSO, fans out to Yelp, TripAdvisor, etc., fans out to top lists in well-known publications (Times Union, ILOVENY, etc.), and finds top lists there. Then it syncs all of it in one answer with notes on how all those places are different, what to pay attention to, and what others said about each place.
Now, let’s take a more specific query: “I want to eat fish tacos today, but I prefer cod over other fish. Find me a restaurant that serves those nearby” (which is more likely to happen when users are interacting with LLM platforms.)
Google search: Pulls map results of all businesses mentioning “fish tacos” in the description, on the site, or in the reviews.
ChatGPT/Gemini: Pulls map results of all businesses mentioning “fish tacos” in the description, on the site, or in the reviews. BUT ALSO, fans out to Yelp, TripAdvisor, etc., fans out to top lists in well-known publications (Times Union, ILOVENY, etc.), and finds top lists there. THEN ChatGPT will read the menus, reviews, etc., to make sure there are, in fact, cod tacos served there. And it will only give you options where it can find the actual cod tacos.
Unless your site, reviews, or lists mention that you are solving a specific problem, you won’t be mentioned or cited in AI Answers
Knowing all of that, here are additional (GEO?) tactics you need to keep in mind when optimizing for AI, apart from traditional SEO:
Being part of LMM training data will help your brand appear in AI answers more consistently, but it requires a strategic approach that goes beyond traditional SEO tactics. You need to create a tightly relevant, consistent footprint on your site and in external mentions (product reviews, forums, etc.).
Being indexed by both Google and Bing is mostly fundamental for being found in AI answers, but there are additional efforts involved, like making sure your site is easy to access and crawl by AI bots, structuring your content to make it easier for AI to find answers in, and optimizing for specific problems/prompts in addition to optimizing for keywords.
Keyword optimization is very important for both, but for AI platforms, you also need to keep “fan-out” queries in mind.
One of the most common questions here is:
Isn’t SEO Enough?
In most cases, organic SEO visibility will result in better AI visibility, whether directly or indirectly:
Some AI solutions do rely on Google rankings more than others (AI Overviews, for example, work by summarizing ranking URLs for the current and fan-out queries). Ranking higher in Google will directly result in better visibility in AI Overviews (and also Perplexity).
Google is still the most powerful visibility engine. If your site is visible in organic search, your brand is better known (as more people find it, discuss it online, etc.). Relevant brand awareness drives better visibility in AI Answers. This way, organic rankings drive AI “rankings” indirectly, as it is just a better-known brand.
So at this point, a well-planned, consistent organic search optimization strategy is a good way to gain visibility in generative AI Answers.
So Do We Need GEO?
Google remains the biggest search engine, and it is also becoming one of the AI leaders. It does look like organic search visibility is not only fundamental but also key to AI visibility at this point.
And yet, additional tactics are already helpful to increase that visibility and better prepare for the future:
A more strategic approach to building a tightly relevant footprint can help your brand get better positioned in training data (so your brand will be part of the answer even when AI doesn’t perform a search).
Adjusting content to AI crawlers (through a better structure and possibly schema) will help your URLs get cited more throughout more platforms.
GEO is not fundamentally different from SEO, and they often overlap. It is still all about facilitating relationships between a digital brand (website) and buyers (human beings for now, but also AI agents in the future). In essence, GEO is a branch of an SEO strategy, accommodating AI technology better by keeping in mind its limitations and specifics.
If you need help with either or both, I am one free call away ;)