Prompt language
原創 idea 成本高又花時間。識賺錢嘅人，係做市場已經證明有需求、但又唔同人正面競爭嘅嘢。App Store 就係你嘅市場研究工具：搵一個冇買廣告都排第一嘅 App，佢啲負評，就係用戶未被滿足嘅需求。
📸 Step 1｜準備素材
- 揀你有興趣嘅範疇，去 App Store search。
- 開第一個唔係 Sponsored 嘅 App。冇買廣告都排第一，即係真係好多人搵。
- Screenshot 三樣嘢：
- App 簡介：成頁，包括功能列表
- 好評：10 條左右
- 負評：至少 15–20 條（呢啲最值錢）
負評影得夠多，個分析先會準。 AI 冇料就會靠估。
🎯 Step 2｜主 Prompt
將所有截圖丟入 Claude，然後貼上：
Prompt 1：逆向分析
Reverse engineer this app for me, and work out: - Who is its ideal target customer? - What is that customer's biggest pain point? - What gap in the market does this app still fail to fill? - If we added AI to make this app easier and more convenient to use, which AI features would we add? I want to focus on serving one core ideal customer profile (ICP) and take market share from this app.
🔬 Step 3｜追問（唔好慳呢步）
AI 第一次答通常太靚仔。跟住問返佢：
Prompt 2：追問
For every gap and AI feature you proposed, evaluate it again: 1. Which reviews is this conclusion based on? Quote them. 2. If the pain point is really this big, why hasn't the original app addressed it? A technical limit, a conflict with its business model, or simple neglect? 3. Does this AI feature genuinely solve the pain point, or is it just "adding AI"? 4. What evidence is there that my chosen ICP would pay for this? Wherever you are not sure, say "this needs to be validated with real users" instead of guessing.
睇下佢引唔引到原文。 引唔到嘅結論，就當係佢作嘅。
✅ Step 4｜問返真實用戶
AI 畀你嘅係起點，唔係答案。搵 5 個目標用戶，問呢 5 條：
- 你依家用緊咩解決呢件事？
- 上一次用嘅時候，最煩係邊部分？
- 你試過搵咩代替？點解冇繼續用？
- 如果有樣嘢幫你解決到，你會願意畀幾錢？
- 邊一刻你會諗起要用呢樣嘢？
唔好問「你會唔會用」。 所有人都答會。只問佢做過咩。
做到呢度，你手上有一個缺口，同埋 5 個真實用戶嘅答案。 下一步係：驗證佢真唔真、低成本整出嚟、同埋賣得出。 呢三步，先係大部分人卡住嘅地方。
