Schema.org markup and entity recognition in 2026

Treating the site as data, not just pages

In 2026, the SEO work that mattered most for Aftertype was treating the site as data instead of just a collection of pages. Concretely, that meant site-wide schema.org markup and a companion llms.txt file, so both classic search engines and AI systems could understand what the site is, not just what it says.

From impressions to entity recognition

Before structured data, Search Console made Aftertype look like a pile of pages and keywords. Impressions went up and down, clicks were fine, and the query mix was a foggy blend of terms that roughly matched the content.

After rolling out structured data across the site, the interesting change was how the site appeared inside Google’s own model of the world. Queries started clustering around Aftertype as an entity, not just as text on a page. Search Console made it clearer that the site was being treated as a thing in the graph, not just a set of strings.

What actually went into the markup

The work was not magic. It was mostly tedious, careful markup across three main types: Organization, Article, and FAQPage.

Organization schema

First step: describe Aftertype as an Organization. That meant being explicit about:

• Official name
• URL and logo
• Contact details
• sameAs links to profiles elsewhere

This is the base layer. Without it, every other piece of content floats around without an owner. With it, articles and FAQs can all point back to a single, stable entity.

Article schema across the site

Next step: treat each significant piece of content as an Article instead of a generic page. That meant marking up:

• headline and description
• author, publisher, and datePublished
• mainEntityOfPage to connect the article to the canonical URL

This is boring, mechanical work, but it gives search engines a clean graph: this organization publishes these articles, on these topics, at these times.

FAQPage schema where it actually fit

FAQPage schema only went on content that was genuinely question and answer shaped. Each question and answer pair was marked up as Question and Answer inside a FAQPage.

The goal was simple: give both search engines and AI systems a clean, structured set of canonical answers for recurring questions, instead of scattering them across loosely formatted paragraphs.

llms.txt: a companion for AI search

Alongside the Schema.org work, Aftertype added a companion llms.txt file. Classic search engines read robots.txt to understand what they can crawl. AI-focused crawlers and LLM-backed search tools need a similar pointer, but with extra hints about what is actually useful for training or retrieval.

Where Schema.org describes the structure of each page, llms.txt describes which parts of the site are high signal, canonical, or stable enough to be worth ingesting. Together, they act as two halves of the same idea: one for traditional SEO, one for AI search.

GEO vs SEO in 2026

By 2026, the line between SEO and what you might call GEO (graph or entity optimization) is very visible.

• SEO is still about ranking pages for queries.
• GEO is about making sure you, your company, and your content exist as clear entities in the graph that search engines and LLMs use.

The Schema.org and llms.txt work for Aftertype sat squarely in the GEO bucket. Instead of chasing another round of on-page tweaks, the focus was on telling machines:

• “This is the organization behind this site.”
• “These are the articles and what they are about.”
• “These are the canonical answers to recurring questions.”

Once that graph was in place, traditional SEO metrics started to look different. Branded and related queries clustered more logically, and Search Console began to look less like noise and more like a map of how the entity was understood.

What changed in Search Console

After deploying structured data site-wide on Aftertype, Search Console results shifted in a few ways that matched the GEO focus:

• Queries tied more clearly to the organization name and core topics, instead of random long-tail matches.
• Branded and entity-like queries became easier to track as a coherent group.
• Rich results and enhanced snippets aligned more cleanly with marked-up FAQs and articles.

This was not about a viral spike or overnight doubling of traffic. It was a change in how the site was categorized and presented, and in how new content slotted into an existing entity graph instead of starting from zero each time.

Why this beat other SEO habits in 2026

Most SEO activity in 2026 feels like tuning a radio station that keeps changing frequency. Algorithms shift, SERP layouts change, and AI overviews eat more of the page. The Schema.org and llms.txt work for Aftertype targeted a slower-moving layer: how the site is modeled as data.

Instead of guessing which micro-optimizations might please the latest search experiment, the work focused on:

• Making the organization unambiguous as an entity
• Making each article unambiguous as a piece of content
• Making FAQs unambiguous as canonical answers
• Making AI crawlers unambiguous about what to read and reuse

That combination turned a loose collection of pages into something closer to an API for both search engines and LLMs. The payoff showed up less as raw volume and more as recognition.

If you only do one thing

If you are deciding where to spend limited SEO time in 2026, a simple order of operations looks like this:

1. Implement solid Organization schema site-wide.
2. Mark every significant content piece as an Article with proper metadata.
3. Use FAQPage schema only where the content is truly question and answer shaped.
4. Add an llms.txt file that points AI crawlers at your best, most stable content.

Everything else can wait. Entity recognition is the part that actually sticks.