The six sessions 💬 · one message, rebuilt in all of them
You are Daybreak's data lead and the subscription redesign is three weeks late. You have to tell the founder. That is the message. It is currently 163 words, it contains nine corporate abstractions, and the thing you want from them is in the last sentence, which is why the last three updates got no reply.
The ask
Who is reading it, what you want them to actually do, and the one-sentence test. Most updates that get no reply contain no ask at all, and their authors have never noticed.
🎧 Session 2 · easyListening first
You cannot structure a message for somebody you have not heard. The two questions to ask before drafting, what a non-reply is telling you, and reflecting back as a correction mechanism.
🧭 Session 3 · mediumStructure
The ask goes first. Three skeletons worth knowing, the buried-lede failure with a number on it, and why the same structure survives an email, a document and a room.
📝 Session 4 · mediumThe written version
The scorer. Sentence length, passive voice, abstraction, specifics, and ask position - all measured on your real text. Take the message from 5 to 100, then switch on the lever everybody asks for.
🎤 Session 5 · hands-onThe spoken version
The same skeleton out loud. The opening line, what has to be cut, why the interruption is usually the real question, and using silence on purpose.
🪨 Session 6 · hands-onBad news and disagreement
Delivering it upward early and specifically, disagreeing without damage, taking the pushback without defending, and the repair conversation nobody teaches.
Your passport 🎫
Clear all three quiz questions in a session and it gets stamped. Six stamps means one message taken from ignored to acted on. Stamps live in this browser only.
What this course is not, so you land in the right place 🧭
| If your question is | Go here instead |
|---|---|
| How do I use AI to draft the writing that fills my week? | learn-ai-writing-with-phoebe |
| How do I document software properly? | learn-tech-writing-with-phoebe |
| How do I build a deck that carries a talk? | learn-ai-ppt-with-phoebe |
| How do I design learning that changes behaviour? | learn-course-design-with-phoebe |
| How do I say a thing to a person so that they act on it? | you are in the right place |
Deliberately AI-agnostic. Every judgement in here - who the reader is, what you actually want, what to cut, when to stop talking - is one a drafting tool cannot make for you.
Choose your path 🗺
Three doors into the same conviction: being understood is a mechanical skill with measurable parts, not a personality you either have or do not.
What is honest here 🧾
The scorer is mostly measurement and partly a rubric, and it says which is which on its own footer.
| Claim | Status |
|---|---|
| Where your ask is, sentence N of M | Really measured on your own text. It finds the first sentence that asks for a decision or an action and reports its position |
| Words per sentence, specifics, total length | Really measured. Plain string analysis, no model involved |
| Passive-voice count | Measured by a regex heuristic over a fixed participle list. It will miss some and occasionally flag a false one. The page says so where the number appears |
| Abstraction count | Measured against a list of 30 entries, chosen not derived. A starting point for noticing, not a standard |
| The score out of 100 | A stated rubric weighting those six measurements. Two people can write a 70 and an 85 and the 70 can be the better message for its reader |
| Daybreak, the founder, the message | A teaching scenario. The measurement code is real; the situation is written |
The knowledge map 🧠
The whole course at a glance - hover a session to spotlight its concepts, click any node to jump in.