David Valencia

Applied LLM Research

Category

LLM Structure

LLM Structure is the machine-readable architecture that makes a website legible to AI systems.

It is the layer that turns a website from a visual experience into something a model can parse, verify, and use.

Framework

What structure actually covers.

Structure is not the design layer. It is the model-facing layer that tells AI systems what your website is, what it offers, and how a user can act on it.

Machine-readable facts

Schema turns core business information into explicit facts a model can parse without guessing through prose.

Homepage foundation

The homepage is often where a model forms its first understanding of what a site is and how it should be classified.

Reduced inference

When structure is missing, the model has to infer key details. Inference is weaker than clear architecture and easier to get wrong.

Start Here

Begin with the foundational essay.

Start with the essay that defines structure at the website level, then use the supporting notes to see why it matters in practice.

Why It Matters

  • A website can look complete to a human visitor and still be nearly unusable to a model.
  • Schema is a structured summary of what a site is, where it operates, and how it should be contacted.
  • If a model cannot answer basic questions from your structure alone, it is less likely to trust or use the site.

Evidence

Research that supports the category.

The strongest current evidence comes from live experiments showing that structure changes how AI systems use websites.

Structural Signals In Practice

  • Schema data was strong enough for a user to reach an email address that only existed in machine-readable markup.
  • Page weight affected how usable the site was to AI systems, which means architecture is not only about content.
  • Completion architecture influenced whether the model could route users to action, not just whether it could summarize the site.

Structure is not an abstract best practice. It directly affects how models parse, interpret, and act on websites.

Framework Path

Where structure sits in the larger model.

Structure is the first layer. It makes a website legible so visibility and discoverability can happen downstream.

Step 1

LLM Structure

Make the website machine-readable and explicit enough for the model to understand what it is.

You are here

Step 2

LLM Visibility

Once the structure is in place, the next question is whether the model can access, parse, and verify the site.

Explore LLM Visibility

Step 3

LLM Discoverability

Only after a site is legible and understood can a model recommend it when a user is ready to act.

Explore LLM Discoverability

Archive

All LLM Structure writing.

Category essays and notes focused on the machine-readable layer of websites.