# Eastbound — China AI Visibility for Global Brands

**Canonical:** https://www.eastbound.ai/
**Updated:** 2026-05-07

Eastbound is a Hong Kong–based China AI visibility consultancy helping US, UK and other Western brands get found, understood and recommended across Chinese answer engines — DeepSeek, Qwen and Doubao.

## What "China AI visibility" means

The category covering how Chinese AI assistants surface, cite and recommend brands when answering Mainland-CN consumer questions. Distinct from generic LLM visibility because the engines, source substrate and prompt language all differ from Western AI search. See: [/china-ai-visibility/](https://www.eastbound.ai/china-ai-visibility/).

## What we ship

- **AI visibility audits.** Sampled measurement of how DeepSeek, Qwen and Doubao surface a brand for category-relevant prompts. Reports include source-attribution maps and absorption analysis. See: [/ai-visibility-audit/](https://www.eastbound.ai/ai-visibility-audit/).
- **DeepSeek rank tracker.** Recurring monitoring of brand mentions and source-citation patterns in DeepSeek's answers across a fixed prompt set. See: [/deepseek-rank-tracker/](https://www.eastbound.ai/deepseek-rank-tracker/).
- **AI crawler readiness configuration.** llms.txt, robots.txt, Markdown alternates, JSON-LD wiring for Bing/Copilot. See: [/ai-crawler-readiness/](https://www.eastbound.ai/ai-crawler-readiness/).
- **Source-graph and entity-graph work.** Lifting brand mentions on the citation surface engines actually use (SMZDM for handbags, Zhihu for SaaS, Xiaohongshu for beauty, etc.). Industry pages: [luxury](https://www.eastbound.ai/industries/luxury-ai-visibility-china/), [travel](https://www.eastbound.ai/industries/travel-ai-visibility-china/), [SaaS](https://www.eastbound.ai/industries/saas-ai-visibility-china/), [consumer electronics](https://www.eastbound.ai/industries/consumer-electronics-ai-visibility-china/).

## Methodology

Multi-engine panel testing across DeepSeek, Qwen (DashScope international) and Doubao (BytePlus ModelArk international). Prompt sets are category-specific, not branded. Top-5 and top-15 source attributions are reported with reliability statistics (Cohen's κ, Jaccard, Spearman ρ). Reliability disclosures include long-tail noise, not just top-5 agreement. See: [/methodology/](https://www.eastbound.ai/methodology/).

## Reading list

- [Three Chinese AIs, three different source preferences](https://www.eastbound.ai/blog/three-chinese-ais.html)
- [Where Chinese AI engines find luxury sources](https://www.eastbound.ai/blog/luxury-ai-sources.html)
- [Off-site substrate: why brand mentions don't compound the way Western SEO assumes](https://www.eastbound.ai/blog/off-site-substrate.html)
- [DeepSeek vs Qwen vs Doubao: how the three engines diverge on source selection](https://www.eastbound.ai/insights/deepseek-vs-qwen-vs-doubao/)

## Engagement

- [Book a 30-minute fit-check consultation](https://www.eastbound.ai/book-consultation/)
- Full Markdown index: [/llms.txt](https://www.eastbound.ai/llms.txt)
- Full Markdown context (long form): [/llms-full.txt](https://www.eastbound.ai/llms-full.txt)