Query Fan-Out Content Strategy (AEO/GEO)
Win AI answers by understanding that a machine, not a person, is doing the searching behind every response
- Difficulty
- Moderate
- Time to result
- ~months to results
- Steps
- 3
- Confidence
- 88%
A model of how AI search actually retrieves and cites content, plus what creators should do about it. Use it when deciding how to make your content show up in AI Overviews, AI Mode, and other AI answer engines instead of chasing SEO tricks.
Origin
Robby Stein explained the under-the-hood mechanics of Google's AI search when asked about the rise of AEO/GEO (the AI-era evolution of SEO). Rather than give a vague 'make good stuff' answer, he described query fan-out and tied it back to Google's decades-old human rater guidelines.
Core principles
- 01When AI constructs a response it runs 'query fan-out' — appending dozens of background queries and using Google search as a tool, so something IS still searching, just not a person.
- 02Each background search is paired with content, so classic quality signals still decide whether your page is used.
- 03AI use is expansionary and skews toward advice, how-to, and complex needs — so target the content types people actually turn to AI for.
How to run it
- 1
Model the fan-out, not the single query
Understand that for one user question the model appends and runs dozens of related queries in the background, then pairs each with content. You are competing to be the content matched to those sub-queries, not one keyword.
Pro tip Read Google's public human rater (Search Quality) guidelines — Stein points to them as the explicit rubric for what counts as great information.
- 2
Satisfy intent, source, and be original
Make each page genuinely helpful for a specific intent: satisfy what the user is trying to get, cite your information, and be original rather than repeating what has been said 500 times.
Watch out Repeating widely-repeated content is a negative signal — originality and citations are what get you linked and quoted.
- 3
Aim at what people use AI for
Because AI use is expansionary and advice/how-to/complex-need heavy, ask: what kind of content is someone using AI for, and how can my content be the best for that specific set of needs?
Pro tip Pick the need where AI demand is growing (advice, planning, comparisons) and be the single best resource for it rather than covering simple lookups.
In the wild
Stein described asking about specific shoes: the model appends dozens of queries and searches Google's backend in the background, pulling real-time data, then pairs each search with content to build the answer.
→ Pages designed to be extremely helpful for those sub-queries are the ones surfaced and cited in the response.
Common mistakes
Treating AEO as a new trick separate from quality
Stein says the core signals — does this satisfy intent, is it sourced, is it original — are 'still extremely valid.' Chasing hacks instead of being the best answer for a need misreads how the retrieval works.
Is it for you?
Best for
Content creators, SEO/marketing teams, and founders who need their pages surfaced inside AI answers
Not ideal for
Brands whose value is pure real-time transactional data (phone numbers, prices) where the AI answers directly without linking deep content
From the transcript
“It'll add and append all these other queries like maybe dozens of queries and start start searching basically in the background”
“do you satisfy the user intent of what they're trying to get? Do you have sources?”
“if I were a creator, I would be thinking, what kind of content is someone using AI for? And then how could my content be…”
From the episode
Inside Google's AI turnaround: The rise of AI Mode, strategy behind AI Overviews, and their vision for AI-powered search
Robby Stein (VP of Product, Google Search)