As I discussed on the Tallest Tree SEO Show, there’s a lot of chatter about Generative Engine Optimization (GEO) and AI Optimization (AIO), but what does AI truly mean for your search strategy?
The key is understanding that AI isn’t completely replacing search; it’s adding a new layer on top.
Think of it as a significant UI change, not a fundamental rebuild of the search engine’s core. Google is still crawling, indexing, and using systems like the Knowledge Graph, BERT, and MUM to match user queries with relevant content.
As Eli Schwartz wisely put it in his newsletter, you don’t need a GEO tool or an AIO tool—you need an SEO strategy.
Reinventing the Wheel
I’ve seen articles claiming to know the secret to AI success. Some say there are 10 things to do, other say 18. It’s all listicle nonsense, but for the sake of illustrating why GEO isn’t really a thing, let’s take on the “six pillars of AI visibility” from one such article, which are said to be:
- Citation Readiness
- Answer Alignment
- The Knowledge Graph
- Content Authority
- Technical Access
- Competitive Edge
Let’s take these on one-by-one:
- Citation Readiness: This isn’t new. For over a decade, we’ve optimized for featured snippets by creating quotable, stand-alone sentences. While AI overviews process citations differently (synthesizing an answer then finding supporting passages), the goal of clear, concise information remains the same. The AI is becoming more sophisticated at parsing information, so while quotable passages help, it’s not the sole driver for citations from authoritative sources.
- Answer Alignment: Aligning your content with user questions has been a core SEO strategy since before “People Also Ask” boxes appeared in search results in 2015. If your content directly answers user queries, it’s more likely to be surfaced, whether by traditional search or an AI overview
- The Knowledge Graph: Google’s interconnected database of facts has been around for years, understanding relationships between entities (like Harrison Ford and Indiana Jones). The Knowledge Graph feeds the search algorithm, which then informs AI. Optimizing for it is optimizing for basic search visibility, not a unique AI-specific tactic.
- Content Authority: Authority is the bedrock of Google’s ranking. From the original PageRank algorithm to the more recent EEAT (Experience, Expertise, Authoritativeness, Trustworthiness) guidelines, Google has always prioritized credible sources. AI models themselves don’t understand authority in the same way; they synthesize text. Those authority signals are built into the search algorithm, ensuring the AI draws from reliable content.
- Technical Access: This is simply technical SEO – ensuring your site is crawlable, indexable, and free of issues like broken links or improper redirects. If the search engine can’t access your content, the AI can’t either. It’s a precondition, not an AI-specific optimization.
- Competitive Edge: This is where we might find something slightly different. Being widely mentioned and having high click-through rates are crucial. Google increasingly emphasizes brand mentions alongside links. If your brand is frequently discussed across the web, it becomes part of AI’s training data, potentially leading to more citations. However, even this ties back to fundamental brand building and user satisfaction, which have always been part of a strong SEO strategy.
Debunking Misconceptions: What Not to Do
There are other “new” strategies that are either harmful or a complete waste of time.
Vector Embedding & Cosine Similarity
Some suggest optimizing for how Google quantifies a page’s topic for AI—this is called “vector embedding” which uses the mathematical concept of cosine similarity to turn topics into a number.
However, as Marie Haynes points out, Google’s John Mueller has stated that trying to manipulate vector embedding by flooding a site with specific terms is “literally keyword stuffing,” a practice that has been penalized since the early days of search. Let’s not go down that road again.
LLMs.txt
Then there’s this silliness of creating a duplicate robots.txt file and calling it LLMs.txt. This file would be in Markdown format, so LLMs can “read” it and use it as context for a site.
This fails to consider how the tech stack in AI-powered answer generation works. Most LLMs are not “reading” websites. For the most part, they are consulting a search index and rely on those mechanisms to give them clean, readable documents. Consulting an LLMs instruction file would be far more open to manipulation than just processing web pages as they are.
Given that Google doesn’t support LLMs.txt, it’s unlikely to become a standard. As Search Engine Land reports, Google’s Danny Sullivan has confirmed that “normal SEO works for ranking in AI overviews and LLMs.txt won’t be used.”
Trust the Experts (and Your Own Eyes)
Speaking of Danny Sullivan, he recently reiterated at WordCamp that “Good GEO is just good SEO,” a sentiment echoed by Google’s Gary Illyes.
I realize we should be cautious about listening to Googlers. Google’s messaging can sometimes be contradictory. Recently search head Liz Reid claimed that “relatively stable.” Around the same time Google developer advocate Martin Splitt acknowledged “great decoupling” of impressions and clicks.
But in the case of the GEO vs. SEO debate, reputable SEO experts seem to agree with Google.
Carl Hendy reposted to Sullivan’s WordCamp talk and Lily Ray explained that the proliferation of “nonsense” about AIO and GEO is helped along by the manipulation of LLMs themselves. So we’re seeing short-term, blatant manipulation reinforcing itself.
LinkedIn-famous SEO Mark Williams-Cook humorously illustrated that much of this “GEO” advice is simply “SEO” advice update with a quick find-and-replace.
The Real Impact: Measurement Challenges & Differentiation
But not all GEO analysis out there is junk. Backlinko’s analysis of “SEO vs. GEO” points to two key shifts:
- Citation as a Metric: Citations in AI overviews might become as important as direct traffic in measuring visibility. While “direct quotable responses” are still helpful, the AI’s ability to synthesize means quality and authority are paramount.
- Measurement Challenges: Google Search Console may not provide the granular data we’re used to, making success metrics more akin to brand advertising. Appearing in an AI overview might build brand awareness and influence future user decisions, even without an immediate click.
This scenario, much like Don Draper’s famous “greatest advertising opportunity since the invention of cereal” line, means everyone faces the same challenges.
The “easy clicks” from commodity content are dwindling, so we have to restrategize.
Your Path Forward: Lean into Your Expertise
We need to stop whining about the easy content days gone by. That traffic isn’t coming back. The AI layer is going to capture those simple queries where users don’t doubt the AI’s ability to give them an accurate-enough answer.
Now is the time to differentiate yourself. Lean into what makes you unique:
- Mini Bios: Every article you publish should include a mini-bio explaining the author’s expertise. Dedicated bio pages should then showcase their full professional background.
- Real Expertise: Do your authors have years of experience? Show off their unique educational backgrounds, or significant real-world achievements (e.g., testifying before Congress, appearing on major shows).
- Show, Don’t Just Tell: Don’t just list credentials; link to published articles, books, video clips of debates, or public records of accomplishments.
There is no “elegant hack” for AI visibility. Anything marketed as such is either find-and-replace traditional SEO or a risky LLM manipulation that will eventually lead to penalties.
So instead of chasing shiny new tools or fancy acronyms, build a solid, user-centric strategy based on genuinely good content and your own unique expertise. This is what will see you through the AI wave.