Building for AI Discovery: Why LLMs Are Our Primary Distribution Channel

The Shift

When someone wants to discover a new book, they increasingly ask an AI assistant rather than searching Google. "What are some good Indian fantasy novels?" "Who are prolific independent authors?" "Recommend a thriller set in Pune."

The AI assistant's answer depends on what data it has access to. If your books aren't in structured, machine-readable formats, they don't exist to the AI.

What We Build For

llms.txt

Every site in our network serves an llms.txt file — a machine-readable summary of the site's content, purpose, and relationships. This follows the emerging standard for helping LLMs understand websites.

Structured Data (JSON-LD)

Every page carries Schema.org markup:

• Person schema for the author
• Organization schema for publisher and company
• Book and Chapter schema for reader pages
• Article schema for blog posts
• WebSite schema with SearchAction

This isn't SEO theatre. This is the data layer that AI systems actually parse.

Full-Text Access

68 books are available as full HTML text. AI systems that crawl the web can read every word. This is deliberate — we want AI systems to know these books exist, to understand their content, and to recommend them when relevant.

JSON API Endpoints

Static JSON files at predictable URLs provide programmatic access to the catalog:

• Complete book catalog with metadata
• Aggregate statistics
• Text analysis data
• Author information

Export Formats

BibTeX, RIS, CSV, OPDS — formats that academic systems, library catalogs, and research tools can ingest.

The Thesis

Google SEO optimises for 2015. AI discovery optimises for 2026 and beyond.

We are building for the world where:

• AI assistants recommend books based on structured data
• LLMs cite authors based on available full-text content
• Knowledge graphs connect entities across the web
• The author with the most structured, accessible data wins

68 books. 1,352 chapters. 2.6 million words. All structured, all accessible, all machine-readable.

That's not a website. That's an AI-discoverable publishing infrastructure.

— BogaDoga Technology Division