internal runbook · for Angel only
Maor discovery call · run sheet
給自己看的會議行動手冊・Maor 不會看到
Fri 29 May · 15:00
Bali / Asia·Taipei
60 min · Google Meet
60 min · Google Meet
mindset · the heart of this call
心態校準 · 整場會議的核心
This is a needs meeting, not a proposal meeting.
這是「需求會議」，不是「提案會議」。第一順位是 聽懂他真正的問題，不是急著 demo 或賣方案。
- 成熟的顧問不是「問他想要什麼系統 / 功能」，而是「先釐清想解決什麼問題」，協助規劃最佳方案。
- 妳是問題診斷師，幫他釐清「為什麼」而非「要什麼」。
- Demo 不是「簡報」從頭演到尾。是對話工具 — 先聽完，再挑最打中痛點的演，邊演邊問「你的流程是這樣嗎？這欄位對你有用嗎？」
- 把 demo 變成需求確認，不是單向展示。
70%
聽 · listen
30%
說 · share
key statement · AI framing
關於 AI 的核心定位 · 直接唸出來
I think of AI as a way to accelerate the work and reduce the repetitive load. But the real calibration and judgment still need a human to verify and review.
That's why my approach is to use AI as a partner to surface issues faster, not a replacement for human review. The system is designed to catch errors earlier, but it doesn't remove the need for your team's final sign-off.
什麼時候用：
① 對方問「AI 會不會出錯？」② 講 honest about quality 時 ③ 解釋「妳做什麼、不做什麼」時 ④ 收尾收 boundary 時。
−30 min
Pre-call ritual・會前準備
收尾準備、避免技術出包
Do執行
- 打開 maor-discovery-deck.html（11 sections，左 sidebar nav）在 Chrome 新分頁
- 測一次 screen share 那個分頁
- 挑一個乾淨的 HANDOVER.md 開好備用（VS Code preview 模式）
- 這份 runbook 放第二螢幕 / 手機 / 印出來
- 水。手機靜音。關門。
- 會前 5 分鐘測 Google Meet 鏡頭 + 麥克風 + 錄影按鈕位置
Mental心態
這是妳的甜蜜區。他第二封信自己說了 — 他不要法律 / 估值，他要妳每天在做的 architecture 工作。
妳不是去 audition，是去做 fit-check。雙方都在 assess。
Avoid不做
不要過度準備 playground — 不要把 Claude projects 整理得很完美。真實 > 包裝。他要找的是「真的做過的人」，不是雜誌排版。
00:00–00:05
Warm open + consent・暖場 + 錄影同意
建立連結、設定 agenda、徵詢錄影
Greet招呼
Hi Maor, great to finally see you. How's your week going? (let him answer fully — 30–60 sec, don't rush)
Consent錄影同意
Before we start — would it be okay if I record this just for my own internal notes? Nothing shared, just so I can focus on the conversation rather than scribbling. Of course if you'd rather not, totally fine — I'll just take notes.
Cue提示
等他明確說 yes / no。如果 yes → 按錄影並再說一句「OK recording starts now, thanks」。如果 no → 「Of course, no problem」立刻收。不要含糊帶過。
Frame設框
So here's how I'd love to use the hour. Most of the time I want to spend understanding what you're actually dealing with — your current setup, where it breaks, what success would look like for you. Then towards the end I'll share a few things from how I think about these problems, but only the bits that are actually relevant to what you tell me. And we'll close by figuring out together whether there's a real fit. Does that work?
Cue提示
注意 frame 改了 — 從「我會 demo 給你看」改成「我聽你的，再決定要不要分享」。聽 70%，share 30%。
Remind心態提醒
這時候妳一定會緊張、想開始 demo。抗拒這個衝動。
他願意付錢聊，是因為他覺得妳能看懂他的問題。如果妳一進去就 demo，妳就只是另一個 vendor。如果妳花前面 30 分鐘讓他覺得妳真的懂他在面對什麼，妳就是顧問。
00:05–00:35
Diagnose・問題診斷（30 分鐘，最重要的一段）
釐清「為什麼」而非「要什麼」
Identity身份
妳不是來 demo 的。妳是來診斷他真正的問題。
他自己以為的問題（「Claude 用得不順」）跟真正的問題（「我跟律師之間的資訊傳遞 broken」/「我自己 review 時抓不到對方的洞」）可能不一樣。妳的工作是找出真問題。
聽到他講「我需要 X」時，問「為什麼需要 X？解決什麼？」— 不要直接接「OK 我可以幫你做 X」。
Rule原則
這段妳講 < 30%。記下他用的字（一字不漏）— 等下分享時要用回去。他講長就讓他講。長 = 訊號。沉默不可怕，他會繼續講。
Open先寬
- Walk me through what a typical Claude session looks like for you right now.
- Are you using Claude Projects, or mostly individual chats? What's in each Project if you have them?
- Roughly how many documents — dozens, hundreds?
Narrow再窄
- When it breaks for you — what does "breaking" look like? Forgetting, contradicting itself, missing citations?
- Have you tried any structure already? What didn't work?
- What's the part of this that worries you most — not the most annoying, the part you'd most regret getting wrong?
Stakes關鍵
- What's the deadline pressure? Weeks, months?
- Who else touches this — lawyer, accountant, valuator? How are you handing things between them and Claude?
Why-dig挖「為什麼」
- When you say you want [the exact thing he said] — what would solving that actually unlock for you?
- If we did nothing about this, what would the cost be? Hours? Quality risk? Stress?
- What have you tried so far that didn't work — and why do you think it didn't?
Vision願景
- If we built this together and it worked, what does your week look like 6 weeks from now?
- What does "good enough" look like? At what point would you say "this is working"?
Listen聽什麼
- 他用的原句（"going in circles", "losing context", "version chaos"）— 等等分享時用回去
- 他沒提到的東西 — 那是妳可以補的價值
- 語氣變化 — 哪些他平靜講，哪些他繃緊
- 時間 / 預算訊號 — "we have months" vs "I need this in two weeks"
- 他講 team 時用 "I" 還是 "we"？單人 vs 團隊
Reflect回放
Just to make sure I understand — you're saying [X, Y, Z]. Is that right?
(discovery 中段做 1-2 次 reflect back — 證明妳有聽進去)
(discovery 中段做 1-2 次 reflect back — 證明妳有聽進去)
00:35–00:50
Share if asked・按需分享（不是 demo 簡報）
挑 1-2 個最打中痛點的概念，邊講邊問
Goal目標
Deck 是備用資料，不是 script。目標：他離開時記得 2-3 個 takeaway，不是 11 個 section。
1. Reflect先反映
Let me play back what I think I heard. The core thing seems to be [X — his real problem in your words]. The specific frictions are [A], [B]. Does that match?
→ 等他確認 / 修正後再繼續。
→ 等他確認 / 修正後再繼續。
2. Ask徵求 share
OK that helps. I'd love to share one or two ideas that might be useful for [the specific thing]. Mind if I share my screen briefly?
3. Pick挑 1-2 段
"going in circles"
→ 02 衰減曲線。"This is what I think is actually happening."
"version chaos"
→ 08 Live Brief hub。"I have one of these for every project."
"one big project"
→ 07 four projects pipeline
"find weaknesses"
→ 口頭講 adversarial review（deck 沒這段了）
unsure
→ 05 architecture, ask 「哪一層最痛？」
最多開 1-2 段。每段最多 3 分鐘。每段講完停下來問。不要連續開 3 段。
4. HANDOVER實際展示（如需）
If 要 demo Live Brief 實物：切到 freesoul-website/docs/HANDOVER.md 在 VS Code preview 模式（不要看 raw markdown）。
指 3 段：Current state → Recent decisions → Open questions。最多 90 秒，不要逐字唸。
說："Every session I start by giving Claude this single file. It knows where we left off."
Red lines紅線
- ×連續講 90 秒沒被打斷 → 停。問 "what do you think?"
- ×他打斷問問題 → 關掉 deck，跟著走。對話 > 簡報。
- ×絕對不要從 S01 走到 S11。Deck 是地圖，不是 script。
- ×除非他問，不要主動 demo Claude app / Notion。
00:50–00:55
Fit + pricing・契合度、定價
老實 read · 給 range · 承諾 proposal
Honest老實講
Let me be honest about what I'm hearing.
The pieces I'm confident I can help with are [name them — based on what HE said, in HIS words]. That's the architecture layer.
The part I can't help with — I want to be clear — is the legal and valuation substance. I'm not going to be the person who reviews whether your argument is strong or whether a claim is well-founded. That stays with you and your team.
If that division works for you, I'd love to put together a proposal — scope, options, pricing. I want a couple of days to think it through properly rather than quote off the cuff.
The pieces I'm confident I can help with are [name them — based on what HE said, in HIS words]. That's the architecture layer.
The part I can't help with — I want to be clear — is the legal and valuation substance. I'm not going to be the person who reviews whether your argument is strong or whether a claim is well-founded. That stays with you and your team.
If that division works for you, I'd love to put together a proposal — scope, options, pricing. I want a couple of days to think it through properly rather than quote off the cuff.
Stop停
等他反應。不要填靜默。讓他亮牌。
If price他問價
I structure these as a setup phase plus optional ongoing support. Setup is typically a one-time fee in the USD 4–6K range, ongoing support a monthly retainer around USD 1.5–2.5K. The exact number depends on what we agree on the scope. I'll send a proper proposal with specific options Monday or Tuesday.
Probe探 scope
- Roughly how long do you expect submission preparation to take — weeks or months?
- Would you want me involved through the whole submission cycle, or just architecture setup at the start?
- What does success look like for you 4 weeks in? 8 weeks?
"Is that all?"嫌少
If you need more involvement than that, we can scope a heavier engagement. I just want to start conservatively so you see the value before committing more.
"More?"嫌不夠
Honestly, I'd rather under-promise on substance and over-deliver on architecture. If you want someone to also review the legal arguments, I'd suggest pairing me with a legal AI specialist for that piece. I can help you think about what to look for.
00:55–01:00
Close・收尾
確認 next step、暖收
Close收
Thanks for walking me through all of this — it was genuinely interesting and I have a much clearer picture now. Here's what I'd suggest: I'll put together a proposal by [Monday / Tuesday] with 2 options. You take your time deciding, and if anything comes up between now and then, just message me.
Confirm確認
- Is there anything else you'd want me to consider in the proposal?
- Best way to send it — email, or WhatsApp file?
Last最後
Looking forward to it. Have a great rest of your day.
+15 min
Post-call・會後整理
趁記憶清楚收尾、規劃 proposal
Now立刻
- 趁記憶熱寫 1 頁摘要：他的原句、3 大痛點、timeline、success vision
- 記下 2-3 件妳承諾要放進 proposal 的事
- 記下他講出讓妳意外的東西 — 下次的訊號
24h內當日
- WhatsApp 感謝：「Thanks for the chat — already started thinking through scope. Proposal by [day].」
- 週末空 2 小時寫 proposal
- 給 Omri 回報：「Had a great call with Maor, thanks for the intro. Sending him a proposal next week.」
Proposal提案架構
- 1 段問題 framing（用他的原句）
- What you'll do（scope）— 3 phases
- What you won't do（清楚 boundary）
- 2 options：lean（只 setup） + full（setup + 8 週 retainer）
- 每 option 報價 · timeframe · 付款條件
- Next step（簽 proposal → kickoff 日期）
Log記錄
- /log-decision：「第一個付費客戶 discovery — Maor / Forma Consulting」
- /log-project：如果 proposal 簽了就新增 entry
REF
Quick reference・速查
隨時對照
If he says…
| "going in circles" | → 02 + 06 |
| "Claude forgets" | → 02 curve |
| "one big project" | → 07 pipeline |
| "version chaos" | → 08 Live Brief |
| "find weaknesses" | → 口頭講（deck 沒了） |
| "re-explain" | → 08 Live Brief |
| "how long?" | → 不報時間，看 data 量決定 |
| "how much?" | → pricing card ↓ |
Your boundaries · 妳的邊界
| 我做 | workflow, architecture, prompts, training |
| 我不做 | legal substance, valuation correctness |
| 我不看 | discovery call 不看他機密文件 |
| 我會做 | 寄 proposal 給 options，不當場 commit |
| 不確定 | "I want a couple days to think it through" |
pricing card · keep visible
If he asks, you have an answer.
他問價，妳手邊有答案。不報死，給 range，承諾下週寄 proposal。
| Component | Range (USD) | What it covers · 內容 |
|---|---|---|
| Setup phase | $4,000 – $6,000 | 2-3 週。Audit 現況、設計 architecture、配 4 個 Projects、寫 system prompts、建 Live Brief 模板、訓練他自己跑。 |
| Retainer | $1,500 – $2,500 / mo | 每週 check-in、迭代、heavy week 即時支援。至少 2 個月起，之後 month-to-month。 |
| Hourly | $150 / hr | Retainer 範圍外的臨時工作。記時、月結。 |
不要在 call 上報死單一數字。給 range，承諾下週寄正式 proposal。
proposal scope · keep for after-call
What I'll do for him · the 6-step deliverable
這 6 點是 proposal 的 scope，discovery call 上不出示，會後寫 proposal 用。
| # | Step | What it actually means · 內容 |
|---|---|---|
| 01 | Audit your evidence | 看他現有文件、分 stable（reference shelf）vs active（working set）。Map 哪些是長期 reference、哪些是當前要處理的。 |
| 02 | Design the four projects | 決定每個 Claude project 的邊界、輸入、輸出。Claim Inventory / Opposing Counsel / Drafting / Final QA 各自的 scope。 |
| 03 | Write the system prompts | 每個 project 的 instructions。用他真實 case material 測試 + 迭代。 |
| 04 | Build the Live Brief template | 用他團隊的語言、實際 workflow 設計 Live Brief 格式。 |
| 05 | Train you to run it | 2-3 次 session 走一遍 real case，直到他自己可以跑、不需要我在場。 |
| 06 | Iterate as you use it | 前幾週實戰，surface 細節 + tuning。retainer 的核心 value。 |
Call 上他若問「妳實際會做什麼？」回："Great question. Let me think it through properly based on what you've told me today, and I'll send a proper breakdown in the proposal. Each will be tuned to your actual situation, not a template."
Mantras
階段之間對自己默念
- 這是需求會議，不是提案會議。
- 我是問題診斷師，不是 vendor。
- 聽 70% / 說 30%。連續講 90 秒沒被打斷 = over-share。
- 他講「我要 X」→ 我問「為什麼 X？解決什麼？」不接「OK 我幫你做 X」。
- Demo 是對話工具，不是簡報。show + ask，不是 show only。
- 不填靜默。沉默是他正在思考。
- Range，不報死。Proposal 下週寄。
- 誠實 under-promise > 急著 over-promise。