AI APPLICATION EXPERIMENT · ONGOING OPERATIONS

“香港布局观察”: one account, two learning loops

A real Xiaohongshu account tests AI operating methods while building an audience, trust and insight into customer needs. Its experience is already used in courses and advisory case material; service delivery and Skill productization are being developed.

2026-10-26 · About two weeks of observation · Operator-reported

Current followers: 84
New followers · 7 days: +14
Highest post views: 2,877
Shares · interaction example: 7

Interaction example: 574 views · 4 comments · 7 shares. Unfollows · 7 days: 0.

Hong Kong audience 31% · Shenzhen audience 19% · Main age group: 35–44

The content loop already in operation

Comments inform a topic → research → draft → publish → review data. A daily 20:00 data check produces a brief that informs the next content cycle.

AI handles repeatable work: Topic research, copy drafts, cover generation, publishing, data monitoring and suggested comment replies.
People own judgment and relationships: Choose topics, check facts, review tone and handle audience interaction and business conversations.

Grow the account. Reuse what it teaches.

Account-to-business loop

Content and interaction → Audience and trust → Specific needs → Service delivery → Feedback and revision

Content and audience growth have been recorded. The path from needs to delivery is still being connected; delivery feedback will improve content and services.

Method-to-product loop

Applied AI research → Account experiments → Methods and cases → Courses, advisory and Skills → Customer feedback

The case is already used with training customers. Packaging the workflow as a Skill is in preparation; future use will inform research and iteration.

Where the work stands

IN OPERATION
Account operations and data reviews
Content production and comment feedback form a cycle while the account continues to build its audience and trust.
IN USE
Course material and advisory cases
Actual methods, revisions and failures support teaching and customer conversations.
IN DEVELOPMENT
Service delivery and Skill productization
Connect specific needs to service delivery. Skill sales are in preparation.

What this cycle changed

Answer a question: In this account's early sample, policy lists received weaker engagement than content with a contrasting perspective. Topic and presentation comparisons continue.
Design for mobile reading: Cover revisions informed a more positive visual tone; simplified Chinese, short sentences and clear sections improved the publishing approach.
Review rejections and rework: Financial-topic rejections and cover rework inform pre-publication checks and human review.

Next: track the needs each piece of content generates, actual delivery, revenue and human effort. Feed customer experience back into the account, courses and tools.

These are early observations from one account. They do not establish AI's causal effect on growth, transferability to other accounts or commercial returns.

Explore AI courses, diagnosis and workflow pilots How practice informs research