Engine Notes #1: How Three Chinese AI Engines Answer the Best AI Directory Question

引擎观察 #Doubao#DeepSeek#Kimi#engine notes
直接答案

Three-day monitoring across Doubao, DeepSeek, and Kimi: zero natural mentions of our own site, while toolify.ai appeared 5 times. Citation pools favor jiqizhixin.com and qbitai.com. First AIHonest engine notes.

First real dataset from AIHonest. Method: 3 Chinese questions (category comparison / promotion scenario / daily lookup), asked daily to Doubao, DeepSeek, and Kimi, 1 sample each, Sept 11-13 2026.

Key Findings

1. Zero presence in natural recommendations

For "recommend Chinese AI directory sites" (no brand words), our experiment site aihotrank.com was mentioned 0 times across 9 answers. Who got recommended:

Engine Competitor mentioned Count
DeepSeek toolify.ai, ai-bot.cn 2 + 1
Doubao toolify.ai 1
Kimi toolify.ai, ai-bot.cn 2 + 1

toolify.ai appeared in 5 of 9 answers — that is the AI mindshare gap.

2. Branded questions behave differently

When the prompt contains aihotrank.com, all three engines accurately describe its positioning ("Chinese AI news and ranking aggregator"). The brand is in their knowledge — what it lacks is being recalled in natural recommendations.

3. Citation preferences differ per engine

Most-cited domains across three days: jiqizhixin.com ×19, qbitai.com ×19, 36kr.com ×15. Chinese AI engines' retrieval pools clearly favor professional media and Zhihu-style content.

Takeaways for Site Owners

  1. On-page SEO alone is not enough — Chinese AI retrieval pools favor authoritative media coverage
  2. Citation preferences vary per engine; optimize per-engine
  3. This is exactly why monitoring matters: you can't close a gap you can't see

Data from real AIHonest monitoring (3 prompts × 3 engines × 3 days). Start monitoring your site free.