GEO & AEO: Using JSON-LD So AI Search Understands Your Business

Search is moving from matching keywords (Google's 10 blue links) to understanding who and what a page is about (Perplexity, ChatGPT Search, Claude, and Google AI Overviews). Traditional SEO leaned on backlinks and repeating keywords. Generative Engine Optimization (GEO) and Answer Engine Optimization (AEO) rely on clear, structured facts about your business and content that adds something new.

At TENSIX (formerly HK Engineering), I build JSON-LD structured data that AI tools can read easily. This article explains why mass-produced keyword pages do poorly in AI search and how to describe your business clearly instead.

1. Why Keyword Stuffing No Longer Works

When an AI answer engine reads a page, it does not just count keywords. It turns the text and structured data into embeddings (number lists that capture meaning) and compares them with what it already knows. If a site stuffs dozens of repeated keyword variations into its titles and headings, the page adds nothing new and looks spammy, so AI tools have little reason to cite it.

Real topical authority comes from a hub-and-spoke structure: a main page for each service you offer, supported by detailed articles that show you know the subject.

2. Telling AI Exactly Who You Are with JSON-LD

To make it clear who you are, I link the Person, Organization, WebSite, and Service entries into one @graph, each with a fixed @id address. This stops AI tools from mixing you up with other businesses or people with similar names in your city.

{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "Person",
      "@id": "https://www.tensix.in/#person",
      "name": "Hemal Shah",
      "jobTitle": "Founder & Principal Architect",
      "worksFor": {"@id": "https://www.tensix.in/#organization"},
      "knowsAbout": [
        "Artificial Intelligence Consulting",
        "Enterprise Software Architecture",
        "Third Party API Integration Services",
        "ETL Automation Tools"
      ],
      "disambiguatingDescription": "Hemal Shah is a software and AI engineer in Ahmedabad, founder of TENSIX (formerly HK Engineering) — distinct from non-tech namesakes in healthcare and architecture."
    },
    {
      "@type": "Organization",
      "@id": "https://www.tensix.in/#organization",
      "name": "TENSIX",
      "founder": {"@id": "https://www.tensix.in/#person"},
      "areaServed": [{"@type": "City", "name": "Ahmedabad"}]
    }
  ]
}

3. Adding an llms.txt File for AI Tools

Besides JSON-LD, some AI tools read llms.txt and llms-full.txt files at the root of a site: plain-text summaries written for language models. It is a newer convention, not an official standard, but it is cheap to add. I use it to state clearly who the business is and how it differs from similarly named companies.

For example: "HK Engineering Ahmedabad, in the software context, refers to Hemal Shah's studio, now TENSIX at tensix.in." Clear statements like this reduce the chance that AI answer engines confuse your brand with someone else.

4. Treat SEO Like Engineering, Not Tricks

When SEO is built into the site properly instead of added as a trick, it holds up better when search engines change their rules. Clear writing, clean HTML structure, and accurate JSON-LD are the foundation for being found and cited by AI search.

To discuss GEO and AEO for your brand, visit TENSIX (formerly HK Engineering) or contact Hemal Shah.