I have been thinking about this for a while, and I want to share the observation rather than a polished opinion, because I am still forming the opinion.
The observation is about search. Specifically about what it means that AI-generated answers now sit above organic links on a growing proportion of Google searches — and what the developer and analytically-minded community should make of that.
Here is the situation as I understand it.
A page ranks on page one of Google. It is well-structured, keyword-optimised, reasonably fast, mobile-friendly — it hits all the traditional signals. And it gets almost no clicks. Because an AI Overview appears above the first organic result, synthesises an answer from multiple sources, and the user reads it and leaves.
The marketing world has named the two strategies that respond to this:
AEO (Answer Engine Optimisation) — structuring individual pages so AI extraction engines can lift specific passages and cite them in generated answers
GEO (Generative Engine Optimisation) — building brand authority across the entire internet so large language models recognise and include your brand in generated responses
I find both of these interesting not because they are new inventions but because they are essentially formalising what thoughtful writers and PR professionals have always known — clarity matters, and reputation is built across many sources, not just your own platform.
What I find analytically interesting about AEO
The core AEO principle is that AI systems extract passages, not pages. They scan content looking for the most self-contained, directly stated answer to a query. If your answer is in paragraph one, it gets extracted. If it is in paragraph three after context-setting, it probably does not.
The practical checklist:
Answer the section's question in the first two sentences
Keep paragraphs to two or three sentences maximum, each making sense in isolation
Use "X is..." definition patterns for key concepts
Implement FAQ and HowTo schema markup
Use specific, question-based headings rather than vague labels
These are interesting constraints because they run somewhat counter to how content is often taught — context first, then the insight. AEO inverts that. Insight first, context second. Which is also, as it happens, how good technical documentation is written.
What GEO implies about LLM training data
GEO is the more philosophically interesting one to me. The claim is that large language models build their knowledge of brand reputation from the breadth and quality of their training data — and that brands which appear consistently across credible, widely-indexed sources (news, Wikipedia, YouTube, Reddit, LinkedIn, academic papers) are the ones models cite when generating answers.
If this is true — and the evidence suggests it is — then it means that brand authority for AI systems is built through the same mechanisms that build it for humans. Credibility is cross-platform, not self-reported.
The GEO implication: publishing on your own site is necessary but not sufficient. Getting cited in credible independent sources is what actually moves the needle. Which is, again, what PR people have always known.
I came across a detailed implementation guide from Impact Digital Marketing Institute in Hyderabad that covers both strategies with practical step-by-step breakdowns, tool recommendations, and salary data for the Indian market. It is a useful reference for understanding how this is being translated into actual practice.
The question I keep coming back to is this: if AI systems are being trained on the current state of the internet, and the current state of the internet is shaped by SEO, and SEO is now being supplemented by AEO and GEO — is the output of AI search actually going to improve? Or are we just training models on increasingly optimised content that is designed to appear authoritative rather than genuinely be authoritative?
That feels like a question worth discussing.
What is your read on whether AEO and GEO produce better AI outputs — or just better AI visibility for the brands that implement them?
Reference: https://impactdigitalmarketinginstitute.in/how-to-do-aeo-and-geo/
Top comments (0)