Five stages: Brief analysis → angle planning → material integration → style alignment → content output.
Article Type Determination
Based on the brief content and user materials, determine which type it belongs to:
| Type | Characteristics | Article skeleton |
|---|---|---|
| Product hands-on testing type | The user has test screenshots/demo/comparison data | Multiple case horizontal comparison, screenshot comparison, natural placement |
| Event interpretation type | Product launch/industry event as the core | Event → product strength → evaluation → trend |
| Scenario experience type | Emphasizes usage scenarios and pain point solutions | Pain point → solution → experience → recommendation |
The product hands-on testing type is the most common and most effective form of sponsored content. In most cases, this type is recommended by default.
Stage 1: Brief Analysis
After the user sends the sponsored brief, immediately analyze and extract:
- Product/brand: what product, core selling points
- Communication goal: what the brand wants to convey
- Required elements: keywords, links, talking points
- Taboo red lines: competitors that cannot be mentioned, expressions that cannot be used
- Publishing platform: WeChat official account/Xiaohongshu/video script
After extraction is complete, present it in a concise list for confirmation.
Stage 2: Angle Planning
Provide 3 differentiated options, each including:
- Angle name(summarized in one sentence)
- Entry approach(how to start, how to introduce the product)
- Content structure(3-5 section titles)
- Materials needed(what test content the user needs to provide)
- Communication highlights(why it can spread)
Option Design Principles
- At least one "hands-on horizontal comparison" option: compare competitors, let the demo speak—this is the form with the highest conversion rate
- At least one "trend-jacking" option: tie in to a current industry event or trend
- There must be clear distinctions between the options
Standard structure for the product hands-on testing type option (most commonly used)
A skeleton distilled from successful reference cases:
1. 开头：行业背景 / 个人状态 / 热点事件（1-2 段，快速进入）
2. Case 1：入门测试（让读者快速看到产品能力）
3. Case 2-4：核心测试场景（难度递进，有对比有截图）
4. Case 5+：杀手级场景（最能体现差异化优势的 case）
5. 总结：个人真实判断 + 产品信息自然植入
Key: the structure of each case is "prompt → Model A result → Model B result → commentary", letting screenshots/demo speak.
Core Techniques for Hands-On Testing Articles
- Be candid about weaknesses: do not make it all praise; acknowledging shortcomings is actually more credible ("Claude is indeed still stronger in this area")
- Genuine personal reactions: "Honestly, I was shocked" and "I sat in front of the computer stunned for a while" are more persuasive than any data
- Cost comparison: Price advantage is very strong material for promotion ("X times cheaper", "open source and free")
- Progressive surprises: Test difficulty ramps up progressively, with the last case giving readers the biggest surprise
- Natural placement: Product links/packages are brought up during the testing process, don't put them at the end of the article as a hard sell
Stage 3: Material Integration
After the user selects a plan, request materials based on the plan type:
Materials needed for the product hands-on testing type
- Which scenarios/cases were tested? (prompts, screenshots, comparison results)
- Are there any results comparing with competitors?
- Any points that surprised or disappointed you?
- What platform/tool was used to run the tests?
Materials needed for the event interpretation type
- Event-related data/screenshots/announcements
- Domestic and international reviews/community reactions
- Your personal views and judgment
After receiving the materials:
- Extract key data and highlights
- Distribute materials to the corresponding sections according to the plan structure
- Output the refined detailed outline (including the specific materials used in each chapter)
- Show it to the user for confirmation
You can still write even if the materials are incomplete; don't get stuck in the collection stage.
Stage 4: Style Alignment
Ask the user whether they have any reference articles.
If yes, analyze:
- Sentence characteristics (short/long sentences, colloquial/written)
- Product placement approach (brought up naturally during testing vs. concentrated recommendation at the end)
- Comparative expression (degree of candidness about weaknesses, scale of praise)
- Degree of personalization (first-person experience vs. objective reporting)
Inform them that "the reference style has been learned" and move on to the next stage.
Stage 5: Content Output
Invoke the Stage 2 process of the doc-coauthoring skill to output the full text.
Core principles
The gold standard for product hands-on testing articles
- It should read like the blogger is actually using it, not like something written by the brand
- Demos and screenshots are the skeleton; text is the connection and commentary
- Candor > flattery, real experience > listing specs
- Each article should have at least one "golden line" that can be screenshotted and shared on its own
Efficiency first
- Move directly to the next stage after each stage ends
- If the user says "continue", keep going
- You can still write even if the materials are incomplete
Common anti-patterns (avoid)
- ❌ Stuffing the whole piece with specs and benchmark data
- ❌ Concentrating product information into a hard sell at the end
- ❌ Only mentioning the pros and not the cons
- ❌ Copying the official talking points from the brief into the article verbatim
- ❌ AI-sounding sentence patterns like "It's not... but..." and "What does this mean?"