This article is part 1/3 of a series discussing why SEO should be a foundation for AI search, how to get started, and how to measure moving forward.
AI search is becoming more popular by the week and marketers are understandably freaking out. No one wants to miss the boat, so the reflexive response has been to scramble - figure out how AI search works, try to learn what LLMs reward, and pour time and budget into showing up in ChatGPT, Perplexity, Gemini, and whatever else launches next quarter.
I think that instinct is half right. AI search matters and it’s going to matter more and more over time so it warrants attention. But a lot of the chatter online is positioning AI search as the new landscape, SEO the old landscape, and implying (if not demanding) that resources should be diverted away from the now irrelevant SEO in the name of chasing AEO (answer engine optimization). This viewpoint is hype-driven, reactionary, and it forfeits a huge opportunity to let SEO be a springboard for your AI search strategy.
Four reasons why SEO is your best starting point for an AI search strategy:
1. Your SEO research is valuable market research you’ve already paid for
If your company has been doing SEO for any length of time, especially if you’ve done it well, you’re sitting on a vetted, organized list of keywords. Ideally they’re grouped by product, service, or audience segment. That’s not just an SEO asset, it’s real market research. It’s evidence of how your customers talk about their problems and how they go looking for solutions.
People have been using search engines for decades. The data you’ve accumulated about how they phrase queries, which ones are high-volume, and which ones convert is hard-earned behavioral data. The fact that a growing share of them are now asking those same questions of an LLM instead of a search engine does not invalidate that research because the underlying user behavior hasn’t changed. A user still comes to a system, asks a question, and expects an answer. The interface is different. The intent isn’t.
So when marketers throw that research out the window and let some AI tool generate prompts for them from scratch based on semantic relevance alone, they’re trading actual market research for a guess. You should be doing the opposite: treating your SEO research as the most reliable input you have into AEO strategy.
2. We don’t have prompt-level data yet
SEO tools have been collecting data on search behavior for the better part of a decade. That’s why we have concepts like search volume, an estimate of how many times users search a specific phrase per month. That single metric has shaped billions of dollars of content strategy because it lets us distinguish between the way you think customers describe your product and the way they actually describe it. That gap is usually wider than marketers expect, and closing it is a big part of what a good SEO strategy does.
AEO has no equivalent. No prompt-volume data. No ranked list of the most common ways people ask an LLM about your category. AI search is simply too new and the platforms are not divulging the data. Worse, getting to that level of fidelity may take a very long time - longer than it took for SEO more than likely (though AI may accelerate the process). Prompts are fundamentally harder to quantify than keywords. Keywords tend to be short and repetitive. Prompts are conversational, full of context, and sprawling. Pinning down the popularity of any single prompt is a much bigger measurement problem than pinning down the popularity of a keyword.
Until the prompt-level data catches up, if it ever does in the way keyword data has, SEO data is the most reliable proxy we have for how popular a given topic, problem, or question actually is. Users are straddling both systems right now. It’s often the same person using Google in the morning and ChatGPT in the afternoon. Assuming a high-volume SEO keyword roughly maps to a high-frequency AI prompt on the same topic isn’t a leap of faith. It’s the safest inference available.
3. LLMs lean on web search, and that means SEO rankings still matter
A lot of people assume LLMs answer purely from their training data. They don’t. When a model can’t answer confidently from what it’s already “learned,” it falls back on retrieval-augmented generation, or RAG, which is a technical way of saying it runs a web search and reads the results. It behaves, in that moment, exactly like a user on a search engine.
And critically, it has to start somewhere. The exact ranking logic is opaque, but it’s reasonable to assume that AI search, especially Google’s, starts at the top of the search results and works its way down. An LLM may ingest more content than a human (because its attention span is essentially infinite), but the pages it prioritizes are still the ones SEO has elevated. Google has already invested years in making sure the top-ranking pages are the most relevant and highest quality. An LLM pointed at the web is logically going to exploit that signal.
So the implication is that if your SEO rankings erode, your presence in LLM-generated answers is likely to erode with them, particularly for the class of queries that trigger RAG. An “AEO result” for many prompts is, under the hood, a series of SEO results.
4. Google is still a major player in AI search
Google made their name in search and their algorithm was their edge. Do you think Google isn’t going to use that edge to differentiate their AI search experience?
It would be extraordinary for Google not to leverage the ranking logic it has spent decades refining. That algorithm is their single biggest competitive advantage in AI search. No other player has equivalent signal on what constitutes a high-quality page or a trustworthy source. Even Bing has historically been accused of copying Google’s results. Google is the default.
So any reasonable bet about where AI search goes should assume Google will bake the same SEO principles (authority, relevance, quality signals, backlinks) directly into their AEO algorithm. And given Google’s market share and user loyalty, they just may continue to set the de facto standard. Other players will probably converge on something similar.
The takeaway
SEO is not the thing you give up in favor of AEO; it’s the thing you pivot from. Your keyword research is your best map of customer intent. Your rankings are your best insurance policy against getting filtered out of LLM answers. Your content and existing on-page work is the foundation AI search is reading from.
The right question isn’t “Should I still do SEO?” It’s “How do I integrate my SEO program in a way that produces AEO outcomes?”
That’s what Part 2 is about.
Next: How to actually pivot your SEO strategy into an AEO strategy, including the intent-group methodology we use to turn keywords into trackable prompts.