Brief generation
The brief is not a summary of information. It is a decision-making act, repeated at each transition in the working day, and what it does at those moments matters more than the information it contains.
Walking in blind
Running a large team is, at its core, a memory problem. You are carrying dozens of threads simultaneously: what is moving in each person's work, what you committed to and when, where the slow-building tensions are, what needs attention this week rather than this month. No one holds all of this reliably. The threads drop. You walk into a one-on-one and reconstruct the last conversation from fragments. You finish a busy Tuesday unable to account for where the afternoon went.
The standard tools do not help with this. A calendar tells you what is scheduled. A task list tells you what is open. Neither tells you which of those things matters today, or what to carry into the two o'clock. That synthesis has to happen somewhere. Most of the time it happens in your head, from scratch, every morning, at the moment when competing demands are already making their case for your attention.
The brief is built to do that work instead.
A decision, not a review
Reading your calendar is a review. Reading a task list is a review. Both tell you what exists. Neither tells you what matters. A brief is different: it makes a judgement call about relevance, and reading it is an act of commitment. You are accepting a view of the day before the day has had a chance to argue with you.
That distinction matters more than it sounds. A leader who starts the morning by opening Slack is reactive before the day has properly begun. Competing signals have already started shaping the agenda. Reading a brief before any of that changes the order of operations. You have settled your priorities before anything else has claimed them.
This is also why the brief reads your one-on-one notes before writing. If you have a one-on-one today, it pulls context from the last session: what came up, what was left open, what was promised. You walk into the conversation already holding the thread. For a leader managing seven or eight direct relationships, that compound knowledge is hard to sustain without a system.
Timing matters for exactly this reason. The brief is most useful before reactivity sets in: over coffee, before email, before the first notification. Once the day has started pulling at you, re-orientation costs more than most people account for. The morning is the one reliable window before that happens.
Across the day
The brief is not a morning-only ritual. It runs across the working day at three distinct moments, each serving a different cognitive purpose.
During working hours a fresh brief can be generated every thirty minutes, each one reflecting the current state: tasks completed, meetings held, notes taken. The priorities at nine are not always the priorities at two. A meeting that resolves a key decision changes what else needs attention. A task that closes changes the shape of what remains. More practically: after a long meeting or a difficult context switch, the cost of reconstructing where you were is real and routinely underestimated. Generating a brief costs almost nothing. Reconstructing your afternoon from memory, while still half-inside the previous conversation, costs considerably more.
The thirty-minute interval is the minimum at which something meaningful has actually changed. Shorter and generation becomes constant background noise. Longer and a busy afternoon can run three meetings without the brief reflecting any of them.
At the end of the day the brief changes function entirely. It is no longer a prioritisation document. Its job is closure. Knowledge work rarely has a clean ending and the transition from working to not-working is often ambiguous. Ambiguity keeps things active overnight.
The EOD brief forces closure by naming what moved. Not what was attempted. What moved. The Zeigarnik effect describes how incomplete tasks persist in working memory; naming completion is one of the more reliable ways to release them. The second half looks ahead: what is on the calendar tomorrow, what needs to be unblocked. Tomorrow has been accounted for. It does not need to stay active overnight.
The EOD brief generates once, after seven in the evening, and does not refresh. Earlier feels premature for most working days. Later and the window between generating and unwinding collapses. Once the day is done, the system stops re-synthesising it.
The audio brief
The audio brief is the logical end of the same argument: orientation should require almost nothing from you. If the written brief lowers the effort of deciding what matters, the audio brief lowers the effort of receiving it.
The first thirty minutes of a working day are rarely conducive to reading. You are commuting, making coffee, moving between places. Listening requires much less of you than reading does. The audio brief is built for that window.
The brief is voiced using a fixed Australian English voice, chosen for clarity at normal listening speed. A short musical intro accompanies every episode. The jingle is not incidental: a consistent audio cue, repeated each morning, functions as a transition signal. The brain handles state transitions efficiently when given clear markers. Pressing play on a familiar format is a different act from sitting down to read. The former asks almost nothing of you. Sitting down to read is itself a small decision.
The text is prepared for speech before it reaches the TTS engine. Markdown is stripped. Times are written out in full. The aim is a news bulletin, not a document being read aloud. A brief that sounds natural gets listened to. One that sounds robotic gets skipped, and a skipped brief is not a brief.
The audio agent runs only on weekdays. A daily format that runs seven days a week erodes the signal it depends on. Habits are built on predictable rhythms, and a brief that turns up on Saturday carries a different weight from one that arrives on Tuesday morning — because on Tuesday you have been waiting for it. Weekday-only is a deliberate trade: you lose weekend coverage to keep the weekday habit intact. Scarcity, here, is a feature.
The file is served via an RSS feed, subscribable in any podcast client. The brief sits alongside whatever else you listen to in the morning. No separate app, no special interface. The lower the friction, the more reliably it gets used.
The archive
Every brief is written to disk when generated. The archive this creates is not a log for humans to read back through. It serves two functions.
The first is generative memory. Each morning brief reads a recent archived brief before writing. This is how the system avoids repeating yesterday: a concern that appeared in the previous brief but has not moved does not need the same weight today.
The second is queryable history. The chat interface can draw on archived briefs when answering questions about recent weeks: what was being prioritised, what was open, what a particular day looked like. A weekly synthesis reads the week's briefs alongside meeting notes and journal entries. Both uses depend on the same property.
A task records that something was open and then that it was closed. A brief records what was actually carrying weight on a given Tuesday afternoon. That is a different kind of record.