Google’s AI search guide: skip llms.txt, keep the SEO basics

Google Search Central published a guide to optimising websites for generative AI features in Google Search in mid-May 2026 and last updated it on 10 July 2026. It is the first time Google has collected its advice on AI Overviews and AI Mode in one official place.

The core message is simple. Google says its AI features sit on top of the same ranking and quality systems as regular Search. The terms AEO and GEO, popular in marketing circles, are treated by Google as SEO. There is no separate optimisation layer for AI.

This matters before anyone pays for a standalone “AEO” package. If a tool does not improve content quality or site structure, it is unlikely to improve visibility in Google’s AI features either.

The most discussed part is a Mythbusting section that answers popular beliefs one by one. First, llms.txt and other files written for AI. Google states that Search does not use them, so creating one neither helps nor hurts rankings. If other AI services read the file, you can still keep one for them.

Second, “chunking” content into small pieces. The guide says this is unnecessary. Google can understand pages that cover several topics, and there is no ideal page length. You also do not need to rewrite in an AI style, add extra long-tail keywords, or seed fake brand mentions across the web, because ranking and spam systems already work against that.

On structured data, the guide says it is not required for generative AI search and there is no special schema.org markup to add. Google still recommends it because it makes pages eligible for rich results, which remain useful in standard results.

The factor the guide stresses most is non-commodity content: pages with first-hand experience and a distinct point of view, rather than summaries of common knowledge or generic tips anyone could write.

On the technical side, pages must be indexed and eligible to show a snippet. The site must be crawlable, JavaScript SEO must be handled properly, page experience should be good, and duplicate content should be reduced. This is the same SEO checklist that never went away.

The guide also notes that Merchant Center feeds and Google Business Profiles help products and local businesses appear in AI responses. For measurement, it points to the Generative AI performance report in Search Console and warns against third-party tools that claim access to Google’s internal metrics.

Confusion arose because, around the same time, TechWyse reported that Lighthouse 13.3.0, released on 7 May 2026, moved its Agentic Browsing category out of experimental status and into the default configuration. That category includes an llms.txt audit.

The detail matters. According to the report, the audit only flags a server error when fetching the file. A missing file (404) is marked not applicable because the file is optional. The category has no 0–100 score, and it also checks WebMCP integration, agent accessibility and layout stability.

In other words, Chrome is preparing tools for a web where AI agents act on behalf of people. That is a separate matter from ranking in Google Search. TechWyse also reports that Google’s John Mueller compared llms.txt to the keywords meta tag, which Search stopped relying on long ago.

In our view, this guidance helps cut unnecessary spending. Many Thai business sites still have basic problems that directly affect AI visibility: prices, promotions or opening hours locked inside images, Thai-language banners with no real text on the page, or sites built with heavy JavaScript that Google cannot fully read.

Two channels are often overlooked: Google Business Profile for shops and clinics with a physical location, and Merchant Center for online stores. Google names both in the guide, and both are faster to fix than rewriting a whole site.

There is also a shortage of Thai-language content based on real experience. The “10 tips” articles repeated across an industry are what Google calls commodity content. Businesses that write from the front line, with real project examples, process steps, typical price ranges or common customer mistakes, have room to stand out.

Before buying any AEO tool, make sure the basics pass. Then invest in content built on your own knowledge. Creating llms.txt is not wrong, but it should not be near the top of the list if Google is the goal.

If time is limited, work in order of impact: indexing and text trapped in images first, then business data on Google, then a longer-term content plan. Track progress in Search Console each month.

• Check in Search Console that key pages are indexed and snippets are not blocked
• Move important text out of images and into real page text
• Update Google Business Profile and Merchant Center to match your website
• Keep structured data for rich results, without expecting a special AI boost
• Write from your business’s real experience instead of generic summaries