Technical GEO: why AI can't see your website

ARTICLE 2 OF 6 IN THE SERIES | UNDERSTANDING GEO

Your content is excellent. Your team has put a lot of work into every page. But if an AI engine can’t access it, read it, and understand what it contains, it will never rank your site. Technical SEO refers to all the factors that make your site machine-readable.

What LLMs Actually Do When They Analyze Your Website

A generative AI engine doesn’t read your site like a human. It explores it using bots that crawl your pages, extract their structure and content, and then send this data to indexing systems that feed the language models.

At every stage of this process, a technical obstacle can exclude your content from consideration, regardless of its quality. A robots.txt file that’s too restrictive, a page that loads too slowly, an unreadable HTML structure, or too much JavaScript on the pages: these are all closed doors that the bot cannot pass through.

The golden rule of technical SEO is simple: if the machine can’t read it, it can’t index it.

The 4 Top Technical Challenges

1. Structured Data (Schema.org): Speaking the Language of Robots

Structured data consists of code tags added to your HTML page that explicitly describe its content. Without them, a web crawler must figure out on its own whether your page is an article, a product page, a company profile, or a customer testimonial. With them, it knows immediately.

For GEO, the most useful schemas are:

Article / Blog Post: identifies editorial content, its author, and its publication date
Organization / Local Business: describes your business, its industry, and its contact information
FAQ Page: explicitly identifies content in a question-and-answer format, which is highly valued by LLMs
Person: identifies an author’s expertise, a direct indicator of E-E-A-T
Breadcrumb List: communicates your site’s hierarchy

This metadata is invisible to the human eye but serves as a direct signaling vocabulary for AI engines. A well-marked-up page is a page that speaks for itself.

2. Indexability: ensuring that crawlers can access

Before they can read your content, crawlers must be able to access it. Several factors can unintentionally block them:

A robots.txt file that excludes entire sections of the site
Noindex tags mistakenly applied to content pages
A missing, incomplete, or unsubmitted sitemap.xml
Orphaned pages (with no inbound links) that are inaccessible to crawlers

A regular indexability audit is the first step in any GEO project. It’s also the most common type of issue and the easiest to fix.

3. Technical Performance: Speed and Stability

Robots, like users, abandon slow-loading pages. Core Web Vitals (LCP, CLS, FID)… performance metrics defined by Google are factored into the evaluation of a website’s quality, including by the systems that power LLMs.

A page that takes more than 3 seconds to load, that visually jumps around as it loads, or whose server responds erratically is perceived as less reliable. This is a negative signal that undermines your technical authority.

4. Semantic Markup: Structuring Content for Clarity

Beyond structured data, the standard HTML markup on your pages should reflect a logical hierarchy:

Only one H1 tag per page, clearly stating the topic
H2 and H3 tags organized into a coherent hierarchy
Alt attributes on all images (LLMs can’t see images, but they read their descriptions)
Clean, readable, and descriptive URLs

These elements form the semantic framework that crawlers use to understand the organization and hierarchy of your information.

An often-overlooked point: the llms.txt file

Since 2025, a new convention has emerged in the GEO ecosystem: the llms.txt file, placed in the root directory of your site (similar to robots.txt). This file provides LLMs with a structured overview of your site, your positioning, your key content, and your reference pages.

While this standard has not yet been universally adopted by all search engines, it is already recognized by several LLMs and serves as a strong signal of intent. Some systems use it as a primary entry point to understand who you are.

What this means in practical terms for your organization

Technical GEO is not an ongoing project; it is a one-time audit followed by compliance efforts and then regular maintenance. The priority actions to be taken within the first 30 days of a GEO initiative are generally:

Crawlability audit and correction of indexing issues
Implementation of Schema.org markup on strategic pages
Verification and optimization of Core Web Vitals
Implementation or revision of the sitemap and robots.txt
Verification of Hn markup on the most important pages

These actions lay the foundation. Without them, optimizations in the other pillars (content, authority, semantics) remain partially ineffective.

Key Takeaways

Technical SEO: Why AI Systems Can’t See Your Site and How to Fix It

If an AI search engine can’t access, read, and understand your content, it will never rank it, no matter how high-quality it is. Technical SEO hinges on four key areas: Schema.org structured data, which allows crawlers to explicitly classify a page’s content; indexability and crawlability, which ensure access to content; page load speed if it takes more than 3 seconds, LLMs may ignore the page; and machine-readable content, meaning clean HTML, consistent Hn tags, and alt text for images.

The golden rule of technical SEO is simple: if a machine can’t read it, it can’t cite it.