The short version
- llms.txt does nothing for Google Search. Google says it ignores these files entirely.
- Structured data is not required for generative AI search, and Google warns against overfocusing on it.
- Content chunking is unnecessary. You do not need to break pages into tiny fragments for AI.
- There is no separate AI ranking system to optimise for. AI features run on core Search ranking.
- There is now a real control: a Search Console toggle to opt out of generative AI features, announced 3 June 2026.
What is GEO, and is it a real discipline?
Generative Engine Optimization (GEO) is the practice of making a page more likely to be surfaced and cited inside AI-generated answers, such as Google AI Overviews, Google AI Mode, ChatGPT search, Claude, and Perplexity. The important nuance in 2026 is that GEO is not a separate technical system you configure. For Google specifically, the AI features are built on the same index and the same ranking systems as ordinary search results, which is why Google's guidance keeps redirecting site owners back to standard SEO fundamentals rather than to new AI-specific files or markup.
Myth 1: "You need an llms.txt file for AI search"
This is the most widely repeated AI SEO recommendation of the last two years, and Google's documentation addresses it directly. It is not ambiguous.
"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."
"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, Guide to Optimizing for Generative AI FeaturesNote the precision of Google's wording: neither harm nor help. An llms.txt file is not dangerous and some non-Google systems may choose to read it, so keeping one is a defensible choice. What is not defensible is selling it as a Google ranking or AI-visibility tactic. If a Nepal agency quotes you for an "AI SEO package" whose main deliverable is an llms.txt file, you are paying for something Google has publicly confirmed it does not read.
Myth 2: "Schema markup gets you into AI Overviews"
"Structured data isn't required for generative AI search."
Google Search Central, Guide to Optimizing for Generative AI FeaturesGoogle also cautions against overfocusing on structured data. This one needs care, because the honest answer is not "schema is useless". Structured data still drives classic rich results: FAQ accordions, review stars, product prices, breadcrumbs, event listings. Those affect how your normal listing looks and how often it gets clicked, which is real value. The correction is narrower than the internet suggests: schema is a rich-results tool, not an AI-citation lever. Keep it, but stop expecting it to buy you AI Overview placements.
Myth 3: "Chunk your content so AI can parse it"
Google's guide states you do not need to break content into tiny pieces for AI to better understand it, and that you do not need to write in a specific way just for generative AI search. What Google does recommend is unremarkable and predates AI entirely: organise content with clear paragraphs, sections and headings, use semantic HTML where feasible, and reduce duplicate content. That is just competent writing and markup. There is no separate AI dialect to learn.
So what actually decides whether AI cites you?
Google's position is that generative AI features are not a separate product bolted onto Search. They run on the same index and the same ranking and quality systems.
"There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary."
Google Search Central, AI Features and Your WebsiteThere are only two hard technical gates, and then one thing that actually matters:
- The page must be indexed and eligible for snippets. If it is blocked by
noindex, or suppressed withnosnippet, it cannot appear. - The site must not be opted out of generative AI features in Search Console.
- The content has to be genuinely worth citing. Google's phrasing is that creating content people find unique, compelling and useful will likely influence presence in generative AI search more than any other suggestion.
Query fan-out: the one mechanic worth understanding
If there is a single technical concept that changes how you plan content in 2026, it is this. Google says AI Overviews and AI Mode use a query fan-out technique: rather than running the user's question as one search, Google issues multiple related searches behind the scenes and assembles an answer from that wider pool. Google says this surfaces "a wider and more diverse set of helpful links" than traditional web search.
The practical consequence is the opposite of keyword-stuffing one phrase. If a page about hiring an SEO expert also genuinely answers what the work involves, what it costs, and how a freelancer differs from an agency, it becomes eligible across several of those hidden sub-queries at once. Depth on adjacent questions beats exact-matching a single head term. This is also why thin pages built around one keyword struggle in AI results even when they rank acceptably in classic search.
The genuinely new thing: Search Console AI controls
On 3 June 2026 Google announced a Search Console toggle that decides whether your site appears in and helps ground responses in generative AI Search features.
"Sites that opt out will not receive traffic or impressions from our generative AI features."
"This control will not be used as a ranking signal for search results outside of these generative AI Search features."
Google, New opportunities, control and insights for website owners, 3 June 2026Alongside it, Google added generative AI performance reporting showing impressions and which pages appear in AI responses, broken down by country. The rollout began with a subset of UK site owners and has been expanding to other countries through 2026, so availability in your own Search Console depends on when your property is included.
For almost every Nepal business the correct setting is opted in, which is the default. Opting out only makes sense for publishers whose business model depends on people landing on the page rather than reading a summary. If you sell services, being summarised with a citation is free distribution.
What this means if you are buying SEO in Nepal
Questionable deliverables
- An "AI SEO package" whose headline item is an llms.txt file
- Charging separately for "GEO" as if it were a distinct technical service
- Rewriting existing pages into fragments for "AI readability"
- Promising AI Overview placements, which nobody can guarantee
Deliverables that hold up
- Getting pages indexed and snippet-eligible in the first place
- Depth across the sub-questions fan-out will ask
- Correct robots.txt for the AI search crawlers, not just Googlebot
- Tracking AI referral traffic so the claim is measurable
One caveat worth stating plainly: everything above is Google's position about Google. ChatGPT, Claude and Perplexity are separate systems with their own crawlers and their own selection logic, and some of them may read files Google ignores. That is exactly why the crawler layer deserves its own treatment, which I cover in the AI crawler and robots.txt guide.
Written by someone who does this for a living
I run SEO, ads and the development behind them
I work with US-based companies on technical SEO and Next.js builds. On one, a 90-day campaign grew organic clicks 271% and AI-assistant referrals 325%.
Frequently Asked Questions
Does llms.txt help SEO?
Not for Google Search. Google states it does not use these files and that maintaining one will neither harm nor help your visibility or rankings, because Google Search ignores them. Other AI systems may choose to read it, so keeping the file is harmless, but it should never be sold as a Google ranking tactic.
Is structured data required to appear in AI Overviews?
No. Google says structured data is not required for generative AI search and warns against overfocusing on it. Schema is still worth implementing for classic rich results such as FAQ, review and breadcrumb snippets, but it is not the mechanism behind AI citations.
Do I need to chunk my content for AI search?
No. Google's guidance says there is no need to break content into tiny pieces, and no need to write in a special way for generative AI. Clear headings, sections and readable paragraphs are recommended, but that is ordinary good writing.
What actually gets a page into AI Overviews?
The page must be indexed and snippet-eligible, and the site must not be opted out of generative AI features in Search Console. Beyond that Google says there are no additional requirements and no special optimizations, because AI features run on core Search ranking. Content that is genuinely unique and useful matters more than any technical tweak.
What is query fan-out?
Query fan-out is Google issuing multiple related searches behind one user question, then assembling the answer from that wider set. It means your page can be cited for a sub-question the user never literally typed, so covering adjacent sub-topics thoroughly matters more than exact-matching a single keyword.
Can I stop my site appearing in AI Overviews?
Yes. A Search Console toggle announced on 3 June 2026 controls whether your site appears in generative AI Search features. Opting out removes traffic and impressions from those features, and Google says it is not used as a ranking signal elsewhere. Page-level control is also available through nosnippet, data-nosnippet and max-snippet.
Sources
- Google's Guide to Optimizing for Generative AI Features on Google Search
- AI Features and Your Website, Google Search Central
- New opportunities, control and insights for website owners (3 June 2026)
Quotations were taken from the live documentation on 23 August 2026. Google updates these pages, so verify the wording if you are citing it later.