Crawlspace

All subjects

generative engine optimization

54 established claims - 8 refused by the grounding gate

Verdict

As of 2026-07-10, Google Search Central's AI optimization guide states that from Google Search's perspective optimizing for generative AI search is still SEO, that many suggested AEO/GEO hacks are not effective or supported by how Google Search actually works, and that machine readable files, AI text files, Markdown, special schema.org markup, and chunking are not needed; it also says llms.txt neither harms nor helps Google Search visibility or rankings, and that no third party tool has access to Google's internal ranking or AI systems. As of 2026-05-20, Google's Lighthouse 13.3 added an Agentic Browsing llms.txt audit while Chrome for Developers' Lighthouse documentation describes the file as an optional emerging convention for LLMs and AI agents, a split that can lead to conflicting instructions with Google's Search docs. Measured findings from 2026 include Ahrefs finding that 97% of llms.txt files across 137,000 domains got zero requests, with the stated caveat that every figure is a ceiling because it measured requests rather than whether bots acted on what they fetched, and SE Ranking finding no connection between having llms.txt and AI citation frequency. The 2024 paper introducing Generative Engine Optimization reports visibility gains up to 40% in generative engine responses but also reports that efficacy varies across domains.

Written from the 54 established claims below, then checked against them. How that works.

Documented position18

What the operator of the system says in its own documentation.

Google Search Central's AI optimization guide states that, from Google Search's perspective, optimizing for generative AI search is…

  • optimizing for the search experience.
  • therefore still SEO.
first-party
1 quote from 1 source
From Google Search's perspective, optimizing for generative AI search is optimizing for the search experience, and thus still SEO.

Google Search Central's AI optimization guide states that many of the 'hacks' suggested under the labels Answer Engine Optimization and Generative Engine Optimization are not…

  • effective.
  • supported by how Google Search actually works.
first-party
1 quote from 1 source
While terms like Answer Engine Optimization (AEO) or Generative Engine Optimization (GEO) are common online, many suggested "hacks" aren't effective or supported by how Google Search actually works.

Google Search Central's AI optimization guide states that creating and maintaining LLMS.txt files will neither harm nor help a site's visibility or rankings in Google Search, because Google Search ignores them.

first-party
1 quote from 1 source
It's completely fine if you decide to create and maintain LLMS.txt files (or other similar files) for other services or systems that use these files. Doing so will neither harm nor help your site's visibility or rankings in Google Search, as Google Search ignores them.

Google Search Central's AI optimization guide states that…

  • no new machine readable files, AI text files, markup or Markdown are needed to appear in Google Search including its generative AI capabilities.
  • Google Search itself does not use machine readable files, AI text files, markup or Markdown.
first-party
1 quote from 1 source
You don't need to create new machine readable files, AI text files, markup, or Markdown to appear in Google Search (including its generative AI capabilities), as Google Search itself doesn't use them.

Google Search Central's AI optimization guide states that…

  • structured data is not required for generative AI search.
  • there is no special schema.org markup that needs to be added.
first-party
1 quote from 1 source
Overfocusing on structured data: Structured data isn't required for generative AI search, and there's no special schema.org markup you need to add.

Google Search Central's AI optimization guide states that there is no requirement to break content into tiny pieces for AI to better understand it.

first-party
1 quote from 1 source
"Chunking" content: There's no requirement to break your content into tiny pieces for AI to better understand it. Google systems are able to understand the nuance of multiple topics on a page and show the relevant piece to users.

Google Search Central's AI optimization guide states that no third-party tool has access to Google's internal ranking or AI systems.

first-party
1 quote from 1 source
Be wary of third-party tools that promise ranking success or claim to use "internal" Google metrics. No third-party tool has access to our internal ranking or AI systems.

Google Search Central's AI optimization guide states that…

  • writing in a specific way just for generative AI search is not necessary.
  • AI systems can understand synonyms and general meanings.
first-party
1 quote from 1 source
Rewriting content just for AI systems: You don't need to write in a specific way just for generative AI search. AI systems can understand synonyms and general meanings of what someone is seeking, in order to connect them with content that might not use the same precise words.

Google Search Central's AI optimization guide states that creating separate content for every possible variation of how people might search is done primarily to manipulate rankings or generative AI responses in Google Search.

first-party
1 quote from 1 source
While it might be tempting to create separate content for every possible variation of how people might search (for example, by focusing on other queries that people have asked, or fan-out queries), doing so primarily to manipulate rankings or generative AI responses in Google Search violates Google's scaled content abuse spam policy.

Google Search Central's AI optimization guide states that seeking inauthentic 'mentions' across the web is not as helpful as it might seem.

first-party
1 quote from 1 source
However, seeking inauthentic "mentions" across the web isn't as helpful as it might seem. Our core ranking systems focus on high-quality content while other systems block spam; our generative AI features depend on both.

For generative engine optimization, Chrome for Developers' Lighthouse documentation describes llms.txt as…

  • an emerging convention.
  • specifically designed for LLMs and AI agents.
first-party
1 quote from 1 source
The llms.txt file is an emerging convention used to provide a machine-readable summary of a website's content, specifically designed for LLMs and AI agents.

Chrome for Developers' Lighthouse documentation states that, for generative engine optimization, the audit is marked Not Applicable if the llms.txt file is not provided by the server.

first-party
1 quote from 1 source
Lighthouse flags the pages if a server error occurs when attempting to retrieve the llms.txt file. If the file is not provided by the server (resulting in a 404), the audit is marked as Not Applicable (N/A), as providing the file is optional at the moment.

Measured finding19

Something someone actually measured, with the method attached.

Search Engine Journal's report on generative engine optimization says…

  • cats.txt is an invented file standard.
  • cats.txt passed all four of the 'proofs' people cite for llms.txt.
  • cats.txt was crawled by the AI bots.
  • Google indexed cats.txt.
  • LLMs returned details about a cat that exists nowhere but cats.txt.
low confidencetrade press
2 quotes from 1 source
so I invented a standard called cats.txt: a text file in which you formally declare your office cats, their jobs, their breeds, and how often they purr.
Then, I checked it against the exact four “proofs” people cite for llms.txt. It passed all four. It was crawled by the AI bots. Google indexed it. LLMs returned details about a cat that exists nowhere but the file.

Search Engine Journal reports that Ahrefs…

  • analyzed logs from 137,000 domains.
  • found that 97% of llms.txt files got zero requests.
medium confidencetrade press
1 quote from 1 source
Ahrefs analyzed logs from 137,000 domains and found 97% of llms.txt files got zero requests. No bots, no humans.
  • Search Engine Journal reports a stated caveat on the Ahrefs llms.txt analysis for generative engine optimization.
  • The stated caveat on the Ahrefs llms.txt analysis for generative engine optimization says every figure in the report is a ceiling.
  • The reason every figure in the Ahrefs llms.txt analysis for generative engine optimization is a ceiling is that Ahrefs measured requests rather than whether bots acted on what they fetched.
medium confidencetrade press
1 quote from 1 source
Every figure in this report is a ceiling. Ahrefs measured requests, not whether bots acted on what they fetched.

According to the Ahrefs llms.txt analysis on generative engine optimization…

  • around 28% of the 137,000 domains publish an llms.txt file.
  • the actual adoption of llms.txt on the broader web is likely lower than 28% because Ahrefs' customers are more technical.
medium confidencetrade press
1 quote from 1 source
The analysis used Ahrefs data to identify user agents fetching files. Around 28% of 137,000 domains publish an llms.txt file, but since Ahrefs’ customers are more technical, the actual adoption on the broader web is likely lower.

In the Ahrefs llms.txt analysis of generative engine optimization, AI retrieval bots linked to ChatGPT and Perplexity made up 1% of the requests to llms.txt files that received any requests.

medium confidencetrade press
1 quote from 1 source
Of files with requests, 96% came from bots, mostly non-AI. AI retrieval bots linked to ChatGPT and Perplexity made up 1%.

Search Engine Journal reports that SE Ranking's earlier analysis…

  • examined 300,000 domains.
  • found no connection between having llms.txt and AI citation frequency.
low confidencetrade press
1 quote from 1 source
SE Ranking’s earlier analysis of 300,000 domains showed no connection between having llms.txt and AI citation frequency.

The academic paper that introduced the term Generative Engine Optimization reports that…

  • GEO can boost visibility by up to 40% in generative engine responses.
  • GEO's visibility boost was evaluated on a benchmark of diverse user queries called GEO-bench.
medium confidencefirst-party
1 quote from 1 source
To address this, we introduce Generative Engine Optimization (GEO), the first novel paradigm to aid content creators in improving their content visibility in generative engine responses through a flexible black-box optimization framework for optimizing and defining visibility metrics. We facilitate systematic evaluation by introducing GEO-bench, a large-scale benchmark of diverse user queries across multiple domains, along with relevant web sources to answer these queries. Through rigorous evaluation, we demonstrate that GEO can boost visibility by up to $40\%$ in generative engine responses.

The paper that introduced Generative Engine Optimization…

  • reports that the efficacy of GEO strategies varies across domains.
  • underscores the need for domain-specific optimization methods.
medium confidencefirst-party
1 quote from 1 source
Moreover, we show the efficacy of these strategies varies across domains, underscoring the need for domain-specific optimization methods.

Practical effect3

What is observed to happen in practice, as distinct from what is documented.

Search Engine Land asserts that agencies which previously sold backlinks are…

  • now pitching "AI mention building" and "LLM visibility campaigns".
  • using the exact same playbook wearing a new mask.
low confidencetrade press
1 quote from 1 source
Agencies that previously sold backlinks are now pitching “AI mention building” and “LLM visibility campaigns.” The sales pitch promises to engineer your brand into ChatGPT, Perplexity, and AI Overviews. Look under the hood, and it’s the exact same playbook wearing a new mask.

Search Engine Journal reports that, in generative engine optimization, the split in Google's guidance on llms.txt can lead to conflicting instructions when comparing Lighthouse or Lighthouse's llms.txt documentation with Google's Search docs.

medium confidencetrade press
1 quote from 1 source
Guidance is split between different Google developer sites, which can lead to conflicting instructions when comparing Lighthouse or its llms.txt documentation with Google’s Search docs.

Changed on4

A dated change. This is where advice quietly stops being true.

  • Google's Lighthouse tool shipped version 13.3.
  • Google's Lighthouse tool version 13.3 added a new Agentic Browsing category.
  • Google's Lighthouse tool's new Agentic Browsing category includes an llms.txt audit.
  • The llms.txt audit in Google's Lighthouse tool's Agentic Browsing category checks whether a site provides the llms.txt file.
medium confidencetrade press
1 quote from 1 source
Days earlier, Google’s Lighthouse tool shipped version 13.3, which added a new Agentic Browsing category. The update includes an llms.txt audit that checks whether a site provides the file and flags server errors when retrieving it.

Disputed between sources2

Sources that do not agree, presented without picking a winner.

Regarding generative engine optimization…

  • Chrome for Developers' Lighthouse documentation instructs site owners to create an llms.txt file and place it in the root directory of their website.
  • Google Search Central's AI optimization guide tells site owners they can ignore creating AI text files like llms.txt for Google Search.
first-party
2 quotes from 2 sources
Create an llms.txt file and place it in the root directory of your website (for example, https://example.com/llms.txt). The file should follow the llms.txt specification and provide a concise Markdown summary of your site's purpose and key links.
Prioritize effective SEO strategies over "AEO/GEO hacks": For Google Search, you can ignore tactics like "chunking" content, creating unnecessary AI text files (like llms.txt), or pursuing inauthentic mentions.

Echoed without a primary source8

Widely repeated in the field, but traceable only to other repetitions. Listed because knowing something is merely echoed is worth as much as knowing it is documented.

A GEO vendor page asserts that…

  • in 2023, researchers from Princeton University, Georgia Tech, the Allen Institute for AI, and IIT Delhi published research titled 'GEO: Generative Engine Optimization'.
  • the research titled 'GEO: Generative Engine Optimization' introduced a GEO-BENCH dataset of 10,000 diverse queries across multiple domains.
low confidenceother
1 quote from 1 source
In 2023, researchers from Princeton University, Georgia Tech, the Allen Institute for AI, and IIT Delhi published groundbreaking research titled "GEO: Generative Engine Optimization". The study introduced the GEO-BENCH dataset, a benchmark of 10,000 diverse queries across multiple domains, and systematically tested nine distinct optimization methods to determine which techniques most effectively improved visibility in AI-generated responses.

A generative engine optimization vendor page asserts that the top three generative engine optimization techniques, Cite Sources, Quotation Addition and Statistics Addition, delivered 30-40% visibility improvements across all content categories.

low confidenceother
1 quote from 1 source
The results revealed dramatic performance differences. The top three techniques—Cite Sources, Quotation Addition, and Statistics Addition—delivered 30-40% visibility improvements across all content categories.

Search Engine Land asserts that Cyrus Shepard found in a recent study that 92% of sites that experience significant organic growth produce their own proprietary assets.

low confidencetrade press
1 quote from 1 source
Cyrus Shepard found in a recent study that 92% of sites that experience significant organic growth produce their own proprietary assets.

According to Search Engine Journal's report on generative engine optimization…

  • ChatGPT confirmed at length that an invented file standard called cats.txt could help its author rank.
  • the author of cats.txt stated that ChatGPT's confirmation is not evidence of anything.
low confidencetrade press
2 quotes from 1 source
I got tired of watching the industry treat “an AI bot fetched it” and “ChatGPT said it helps” as evidence that llms.txt does anything, so I invented a standard called cats.txt: a text file in which you formally declare your office cats, their jobs, their breeds, and how often they purr.
ChatGPT confirmed, at length, that cats.txt could help me rank. None of which is evidence of anything, which was rather the point.

Search Engine Journal states in…

  • one article that Ahrefs ran the numbers on llms.txt across 100,000 domains.
  • another article that Ahrefs analyzed logs from 137,000 domains.
low confidencetrade press
2 quotes from 2 sources
Ahrefs ran the numbers across 100,000 domains and found that the file is, in practice, largely ignored by the crawlers it is meant to court, a finding since echoed by other large studies showing no measurable citation advantage for sites that add one.
Ahrefs analyzed logs from 137,000 domains and found 97% of llms.txt files got zero requests. No bots, no humans.

Refused8

Claims the research produced that the grounding gate would not accept, with its reason. They are here because a site that only shows what passed gives you no way to judge whether the gate does anything.

Show what was refused

Google Search Central's AI optimization guide states that creating separate content for every possible variation of how people might search violates Google's scaled content abuse spam policy.

Refused: The quote only says this violates policy when done primarily to manipulate rankings or generative AI responses, so the claim's unconditional version overreaches.

Unsupported: creating separate content for every possible variation of how people might search violates Google's scaled content abuse spam policy

What it rested on
While it might be tempting to create separate content for every possible variation of how people might search (for example, by focusing on other queries that people have asked, or fan-out queries), doing so primarily to manipulate rankings or generative AI responses in Google Search violates Google's scaled content abuse spam policy.

For generative engine optimization, Chrome for Developers' Lighthouse documentation describes llms.txt as used to provide a machine-readable summary of a website's content.

Refused: The cited quote says llms.txt is designed for LLMs and AI agents, not generative engine optimization.

Unsupported: For generative engine optimization

What it rested on
The llms.txt file is an emerging convention used to provide a machine-readable summary of a website's content, specifically designed for LLMs and AI agents.

Chrome for Developers' Lighthouse documentation states that, for generative engine optimization, providing an llms.txt file is optional at the moment.

Refused: The quote states providing the file is optional, but does not mention generative engine optimization.

Unsupported: for generative engine optimization

What it rested on
Lighthouse flags the pages if a server error occurs when attempting to retrieve the llms.txt file. If the file is not provided by the server (resulting in a 404), the audit is marked as Not Applicable (N/A), as providing the file is optional at the moment.

Generative engine optimization has a documented position, reported by Search Engine Roundtable, that is expressed in John Mueller of Google's Bluesky post: "FWIW no AI system currently uses llms.txt."

Refused: The quote shows John Mueller's Bluesky statement as reported by Search Engine Roundtable, but it does not state or clearly entail that this is a documented position about generative engine optimization.

Unsupported: Generative engine optimization has a documented position

What it rested on
Well, John Mueller of Google chimed in and wrote on Bluesky, "FWIW no AI system currently uses llms.txt."

Search Engine Roundtable quotes John Mueller of Google saying that consumer LLMs and chatbots will fetch your pages for training and grounding.

Refused: The quote does not identify the speaker as John Mueller of Google; it only says "He went on to say."

Unsupported: John Mueller of Google

What it rested on
He went on to say, "It's super-obvious if you look at your server logs. The consumer LLMs / chatbots (the ones that SEOs want traffic from) will fetch your pages - for training and grounding, but none of them fetch the llms.txt file. Maybe they will tomorrow? Maybe I'll win in the lottery tomorrow?"

Search Engine Roundtable quotes John Mueller of Google saying that none of the consumer LLMs and chatbots fetch the llms.txt file.

Refused: The quote establishes the statement via Search Engine Roundtable but does not identify the speaker as John Mueller.

Unsupported: John Mueller of Google

What it rested on
He went on to say, "It's super-obvious if you look at your server logs. The consumer LLMs / chatbots (the ones that SEOs want traffic from) will fetch your pages - for training and grounding, but none of them fetch the llms.txt file. Maybe they will tomorrow? Maybe I'll win in the lottery tomorrow?"

Search Engine Land quotes Wil Reynolds asserting that in generative engine optimization, AI is 100x more likely to force you to have receipts to back up your claims than in the old days of ranking on domain authority.

Refused: The quoted assertion does not mention generative engine optimization; the claim inserts a context not stated in the quotes.

Unsupported: in generative engine optimization

What it rested on
Wil Reynolds said it best:
“In the old days you could just muscle your way to the top of Google with zero receipts that you are “top 3” at XYZ thing. If you had enough domain authority you could rank. AI is 100x more likely to force you to have receipts to back up your claims.”

Search Engine Journal reports that, in generative engine optimization, Google's guidance on llms.txt is split between different Google developer sites.

Refused: The cited quote does not mention generative engine optimization; it only states guidance is split between developer sites, so this contextual framing is not supported.

Unsupported: in generative engine optimization

What it rested on
Guidance is split between different Google developer sites, which can lead to conflicting instructions when comparing Lighthouse or its llms.txt documentation with Google’s Search docs.