Nobody disputes that SEO fundamentals still work. Valuable content. Quality backlinks. Domain authority. These aren’t going anywhere.
It’s how people find information, and the infrastructure determining what gets surfaced that has fundamentally shifted.
AI Overviews now appear on 21% of all searches. ChatGPT cites pages ranking at position 21 or lower. Gemini ignores official brand sites in favor of Reddit threads.
So yes, AEO (Answer Engine Optimization) and GEO (Generative Engine Optimization) are buzzwords. But the shift they describe is real.
Table of contents
The backlink authority model just got complicated
One Redditor summed up the skeptic position perfectly: “I can’t see how AEO/GEO meaningfully differs from traditional SEO practices.”
At surface level, both AEO/GEO prioritize content quality authority signals, and both want your brand associated with credible sources.
But the mechanism for establishing authority has changed.
Traditional SEO uses backlinks as votes of confidence. Your site earns links from authoritative domains, Google interprets that as trust; your rankings improve. Linear. Measurable. Well-understood.
GEO flips that model. Authority now comes from brand presence across the web, not just links pointing to your site.
To illustrate why this difference matters, Ahrefs ran a controlled experiment. They created Xarumei, a fictional luxury fashion brand. They built an official website with the “real” brand story. Then they planted detailed fake backstories across Reddit and Medium—no backlinks, just narrative-rich posts about the brand’s supposed Italian heritage and craftsmanship origins.
The result: Gemini and ChatGPT completely ignored the official website. Instead, they confidently repeated the fabricated Reddit backstory as fact. The AI models saw repetition and narrative detail across multiple sources and treated that as truth, even though the official site contradicted it.
That’s how LLMs determine consensus.
Traditional SEO says: “Earn backlinks to establish your site as the source of truth.”
GEO says: “Generate widespread brand mentions to create recognition and association, because repetition signals credibility to AI models.”
When an LLM repeatedly sees your brand mentioned in “best tool for X” discussions alongside established players (even without direct links), it starts grouping you with them.
Google’s #1 ranking doesn’t mean what it used to
If AEO and GEO were just rebranded SEO, the sites ranking #1 on Google would be the ones AI models quote most frequently.
But they aren’t.
A Semrush study found that ChatGPT regularly cites pages buried at position 21 or lower on Google’s results. Not occasionally. Regularly.
AI models use a technique called query fan-out. When you ask a question, the AI doesn’t run one search; it runs several variations of your query in the background, then looks for consensus across many different sources.
We saw this firsthand while building a Chrome extension that extracts the actual queries ChatGPT runs.
When we asked: “What is the best SEO keyword research tool?”, ChatGPT triggered these searches:
- “Best SEO keyword research tools 2025”
- “SEO keyword research tool comparison features pricing”
- “Top keyword research tools reviews”
- “Keyword research software recommendations”
It’s synthesizing answers from multiple result sets, not just the top-ranking page for your exact query.
This creates an opening.
A small niche blog with a perfectly structured comparison table can get cited over a large corporate site with 10,000 backlinks if that corporate site buries the answer under 2,000 words of filler. The AI model is optimizing for answer extraction, not domain authority.
Traditional SEO: If you’re not on page one, your chances of being seen are near zero.
AEO/GEO: If your content provides a more direct and structured answer than the top result, AI will surface it even from position 50.
That’s not replacing SEO. That’s layering a new discoverability mechanism on top of it.
The traffic loss you’re not tracking
Ahrefs’ latest data backs this up. AI Overviews now appear on 21% of keywords, with informational searches driving nearly all of them.
Translation: Your “how-to” and “what-if” articles—the bread and butter of top-of-funnel SEO—are increasingly being answered directly by Google’s AI summary. The user gets the answer. You get nothing.
The traditional SEO playbook says:
Write comprehensive guides. Capture informational intent. Drive traffic to your site where you can convert them.
The new reality: Google answers the question inline. The user never clicks. Your perfectly optimized guide ranks #1 and generates zero traffic because AI Overviews extracted the answer and presented it as Google’s own content.
“If you rely only on traditional SEO and ignore AEO and GEO, you’ll lose out on traffic.”
— Reddit comment
This isn’t speculation. It’s already happening at scale.
What actually works (and what doesn’t)
First, identify the gap between your content and the AI summary. If Google’s AI Overview is answering the query, look at what it’s citing and what it’s missing. Then optimize your content to fill those gaps and increase your chances of being mentioned as a source.
Structure matters more than word count. Tables, lists, direct answers in the first 100 words—these increase extraction probability. Long-form narrative buried under six paragraphs of intro doesn’t.
Second, shift traffic strategy toward content AI Overviews don’t trigger. Comparison pages. Pricing pages. Product reviews with specific use case breakdowns. These still drive clicks because they require human evaluation, rather than just information extraction.
Traditional SEO: You write guides to capture clicks and build top-of-funnel traffic.
AEO/GEO: You optimize for citations within the AI answer itself, knowing your brand gets mentioned even if the user never visits your page. Then you double down on content types that still generate direct traffic.
Third, manage your brand presence across the web like it’s a ranking signal—because it is. Reddit discussions. Quora threads. Industry forums. Medium posts. These aren’t backlink opportunities. They’re AI training data.
If your brand repeatedly appears in high-quality discussions alongside established competitors, AI models start treating you as a peer.
That’s not gaming the system. That’s understanding how the system works now.
The failure modes nobody talks about
Let’s be clear about what doesn’t work.
- Don’t spam Reddit with fake mentions. The Xarumei experiment worked because it was detailed, narrative-rich content that looked organic. Low-effort self-promotion gets flagged by communities and ignored by AI models.
- Don’t abandon traditional SEO. GEO layers on top of SEO fundamentals. If your content is thin, your site structure is broken, and your backlink profile is weak, optimizing for AI extraction won’t save you.
- Don’t obsess over AI visibility for every keyword. Some queries still drive high-intent clicks. If you rank #1 for “enterprise marketing automation platform pricing,” you’re getting traffic even if AI Overviews appear. That’s bottom-funnel intent AI can’t fully satisfy.
- Don’t treat this as a one-time optimization. AI models evolve. Query fan-out patterns change. Citation preferences shift as training data updates. This requires ongoing monitoring, not a one-month audit.
The Reddit skeptics are right to question whether every agency selling “AEO services” actually knows what they’re doing. Most don’t. They’re repackaging SEO fundamentals with new terminology and charging a premium.
But that doesn’t mean the underlying shift isn’t real.
What to do next
If you want to adapt without chasing every buzzword, here’s where to start:
1. Audit your informational content for AI extraction: Pull your top 20 “how-to” and “what-if” articles. Search those exact queries in ChatGPT and Gemini. See which sources get cited and why. Compare your content structure to cited sources. Identify gaps in directness, structure, or detail.
2. Build structured answer formats into new content: Use tables for comparisons. Lead with direct answers in the first paragraph. Break complex processes into numbered steps. Make your content easy for AI models to parse and extract without ambiguity.
3. Track brand mentions across high-authority communities: Set up Google Alerts for your brand name plus “Reddit,” “Quora,” and key industry forums. Monitor where your brand appears in organic discussions. Engage authentically when relevant, but don’t force it. AI models recognize inorganic promotion patterns.
4. Shift content investment toward click-preserving formats: Prioritize comparison pages, pricing breakdowns, and use-case-specific reviews. These content types still generate direct traffic because they require human judgment, rather than just information extraction.
5. Test and measure AI visibility as a distinct metric: Track citation rate in AI responses separately from traditional rankings. If you rank #3 on Google but never get cited by ChatGPT, that’s a signal your content structure needs work for AI extraction.
SEO didn’t die. It evolved.
Yes, “AEO/GEO” is a buzzword. Agencies will oversell it. Consultants will package it as magic and the Reddit skeptics are right to push back on hype.
But the shift is real.
Traditional SEO still matters, quality content still wins, and backlinks still count. But if you’re optimizing exclusively for Google’s traditional algorithm while ignoring how AI models extract, synthesize, and cite information, you’re playing yesterday’s game.
The question isn’t whether AEO/GEO are real. The question is whether you’re adapting to how AI search actually works now.
For more expert insights on optimizing for AEO and GEO, check out CXL’s SEO program and live/on-demand courses on AI search optimization:
- Optimize pages for AI search with GEO/AEO with Steve Toth
- Increase your visibility and revenue from AI-based discovery engines with Alex Birkett
