# BigSearch AI Labs — Research (R&D)

> Controlled studies on AI entity retrieval, AI Readiness (V.C.A.P.), GEO, and AEO.

- Index: https://bigsearchai.com/research

## Studies

- **Sweet 16: AI Readiness of 273 Black-Owned Businesses** — https://bigsearchai.com/research/sweet16
  - llms.txt: https://bigsearchai.com/research/sweet16/llms.txt
  - 89.4% of Black-owned business websites in San Antonio District One are not AI Ready. Sixteen data points on the new digital divide — AI citability, not just internet access.

- **The Aquaria Test: Entity Fragmentation in AI Search** — https://bigsearchai.com/research/aquaria-test
  - llms.txt: https://bigsearchai.com/research/aquaria-test/llms.txt
  - Five AI models. One question. Five different companies. A controlled experiment revealing how frontier AI systems produce irreconcilable answers when no single entity dominates a name token.

- **The Bounceless Test: Entity Collision in AI Search** — https://bigsearchai.com/research/bounceless-test
  - llms.txt: https://bigsearchai.com/research/bounceless-test/llms.txt
  - We fixed the structured data. The AI still got it wrong. A controlled experiment revealing how frontier AI systems override canonical entity declarations with prior statistical memory.

## Frameworks

- **V.C.A.P.** — Visibility (JSON-LD), Citability (llms.txt), Actionability (MCP), Performability (OpenAPI)
- **GEO** — Generative Engine Optimization (machine-readable infrastructure)
- **AEO** — Answer Engine Optimization (content and authority for AI answers)