yuanyuanyuan430/ganhuo-ai-knowledge-commerce

AI Knowledge Commerce Pipeline

Overview

Apply this skill to turn a knowledge-product idea into a repeatable operating system: source selection, source cards, AI knowledge base, writing style, content factory, WeChat Official Account drafts, Knowledge Planet delivery, private-domain follow-up, and weekly review.

Keep the posture growth-oriented: improve source quality, improve content velocity, improve trust, improve conversion, and lower avoidable complaint, copyright, platform, and refund risk. Avoid treating AI rewriting as a magic cleanup layer; treat AI as the production assistant and the human operator as the editor, source owner, and final publisher.

Minimum Input Packet

Collect or infer these fields before building the brief. If the user gives too little context, ask only for the missing fields that would materially change the output; otherwise state assumptions and produce a first draft.

• niche or category;
• target user and current pain;
• current stage: idea, first articles, existing audience, existing paid product, or optimization;
• available sources and their permission status;
• owned experience, test logs, screenshots, cases, or user questions;
• target channels: WeChat Official Account, Knowledge Planet, Feishu Knowledge Base, Enterprise WeChat, Moments, Xiaohongshu, or other;
• paid deliverable and rough price range;
• available tools: Feishu, n8n, RAG tool, WeChat account, analytics, CRM;
• forbidden claims, sensitive topics, and platform constraints.

Workflow

1. Define the offer: niche, user pain, paid deliverable, price ladder, trust proof, and the reason this product is better than scattered free content.
2. Build the source engine: separate owned experience, authorized material, public references, public reports, interviews, tool tests, and user-submitted cases.
3. Create source cards: record URL or contract, author, date, license or permission status, usable scope, key points, risk notes, and what original angle will be added.
4. Build the knowledge base: organize source cards into topic clusters, user questions, templates, examples, objections, glossary, and case notes.
5. Distill the voice: define tone, pacing, examples, taboo words, title style, CTA style, and the human story bank. Improve recognizability without copying another creator's expression.
6. Run the content factory: generate topic options, outline, draft, source-linked claims, tables, CTA, and platform-specific variants.
7. Review before publishing: check source traceability, factual accuracy, personal information, commercial claims, AI-generated media disclosure, and platform fit.
8. Ship to channels: create WeChat Official Account drafts, prepare Knowledge Planet topics, generate Moments/Enterprise WeChat follow-up copy, and keep public release under human approval.
9. Review the loop: record reads, completion, follows, comments, paid conversion, refund signals, questions, and next week's experiments.

Operating Stack

Prefer the smallest stack that can be reproduced:

• Feishu Bitable: master content table, source cards, platform action table, metrics table.
• Feishu Docs or a local Markdown folder: canonical draft body and knowledge notes.
• Dify, AnythingLLM, RAGFlow, Open WebUI, LlamaIndex, or Haystack: knowledge base and RAG layer.
• n8n: schedule, RSS/API intake, status updates, review forms, and draft creation.
• WeChat Official Account API or manual editor: WeChat Official Account draft creation, with final publish approval by a person.
• Knowledge Planet: manual publishing and member operation unless an official API is clearly available.

Route to heavier frameworks only when the simple stack is no longer enough:

• LangGraph or CrewAI: multi-agent drafting, critique, and review loops.
• Postgres: durable audit log when Feishu Bitable is not enough.
• Custom scripts: bulk import, export, or deterministic linting.

Output

Return a compact Knowledge Commerce Pipeline Brief:

• product positioning;
• source map and source-card schema;
• knowledge-base structure;
• publishing workflow;
• WeChat Official Account article template;
• Knowledge Planet weekly operating rhythm;
• automation map;
• review checklist;
• 30/60/90-day rollout;
• metrics and experiment backlog.

Use tables for operating design. Limit the first execution plan to the smallest set that improves output quality or conversion this week.

Growth Copy Rules

Transform hard guarantee language into improvement language:

• Replace "guaranteed income" with "increase conversion probability" or "validate the revenue model".
• Replace "100% AI written" with "AI drafts, human adds experience and reviews".
• Replace "scraped material" with "authorized material, summaries of public sources, own cases, and tested retrospectives".
• Replace "make money automatically" with "improve efficiency in topic selection, writing, delivery, and review".
• Replace "guaranteed viral" with "verifiable actions that improve read completion, saves, shares, and conversion".

Preserve energy and commercial clarity, but remove claims that cannot be proven.

References

Read these only when needed:

• references/checklist.md for the brief template and scoring checklist.
• references/platform-map.md for Feishu, WeChat, Knowledge Planet, and automation boundaries.
• references/source-policy.md for source-card rules and safer source categories.
• references/prompt-pack.md for reusable prompts.

Operating Principles

• Improve the user's business goal without turning the SOP into a loophole playbook.
• Prefer one repeatable publishing loop over a giant theoretical funnel.
• Keep AI-generated output grounded in source cards and human experience.
• Keep final public publishing and refund-sensitive decisions under human control.
• Treat user trust as the compounding asset: better claims, clearer limits, faster delivery, and visible proof.