Skip to main content
The information that reaches the top of a company is, by default, the worst-instrumented information in the building. Filtered through layers of writers managing perception. Arriving late, after the decisions it should have informed are already drifting. Optimized to convey what the writer wants leadership to think, not what leadership needs to know. A CEO reads ten function updates on Sunday night, written by ten different people in ten different shapes, and forms a picture of the company that doesn’t match what’s actually happening on the ground. A board meeting becomes a series of surprises because the pre-read didn’t predict them. An investor update gets written in two hours of guilt the night before it’s due — conveys neither candor nor calibration, and the investor learns to discount it. The unit of leadership decision-making (the call, the bet, the reallocation) depends on a unit of information (the readout, the summary, the memo) that is typically produced under conditions guaranteed to corrupt it. This guide is a specific instance of the workflows pattern. It pairs with the shareable summaries workflow — that page handles per-meeting outputs shaped for different audiences; this page handles recurring leadership artifacts that roll up across many meetings.

A note on this workflow specifically

Leadership readouts are by definition rollups across many meetings. Earmark refines artifacts within a single meeting today; cross-meeting synthesis is not a one-click action. So this workflow has the largest manual assembly step of any in this guide series. What that means in practice: the four template prompts below are real and produce useful drafts. They run on a Customize context document you assemble manually — pasting in the relevant inputs (key meeting artifacts, function lead updates, metrics) before running the template. The drafting step still saves substantial time. The synthesis step is the writer’s curation work. This is the honest tradeoff. Most of this guide is about how to do the manual side well.

Five decisions before your first readout

Audience

Different readouts reach different audiences. The same content shaped for different audiences is a different artifact. Be explicit: If you’re not sure who the audience is, you’re not ready to write.

Cadence

Predictability beats frequency. A weekly digest that lands every Friday morning for 52 weeks is dramatically more valuable than an excellent monthly digest that lands sometimes on the first and sometimes on the third week. Recommended cadences: The cadence is a commitment to the audience. Breaking it is worse than not starting.

Scope

The hardest part of writing for leadership is deciding what not to include. Every readout should have an explicit scope rule:
  • “The weekly leadership digest covers commercial, product, and people developments. It does not cover individual incident detail unless it’s strategic.”
  • “The board pre-read covers KPIs vs. plan, decisions needed, and major risks. It does not cover normal operating updates.”
A scope rule written down once protects every future artifact from drift. Without it, every readout tries to include everything, and signal density collapses.

Durable home

Every series needs a home where the archive lives so it compounds. A Notion page, a Confluence space, a board portal, an investor archive — whichever fits the audience. The home is what makes the series queryable a year later. The Slack/email send is distribution; the home is record.

Ownership

One person owns the writing for each recurring artifact, even if they synthesize inputs from many. Shared ownership fails reliably — the artifact becomes irregular and uneven. The writer doesn’t have to be the most senior person; they have to be the most accountable.

The four templates

Template A — Weekly leadership digest

For the operating leadership team. Read by 5–15 people. Sets the shared picture of the week.
Three things make this template work. The headline is required and singular. If you cannot pick the most important development of the week, you do not yet have a clear enough picture to write the digest. Write the headline first; everything else is downstream. Risks are capped at three. The cap is the discipline. A list of nine risks is functionally a list of zero risks because nobody can act on nine. The “Asks” section is load-bearing. This is where the digest earns its keep. If the digest never produces asks, leadership isn’t using it as a decision tool — and you should question what it’s for.

Template B — Periodic business review

For a function lead reporting to the executive team, monthly or quarterly. Longer, more analytical, heavier on numbers.
The What didn’t work section is the most important part of this template. Most business reviews bury misses or recast them as opportunities. A review that surfaces misses honestly — without spin, with specifics, with what changes — is dramatically more useful, and dramatically more trusted over time, than one that doesn’t.

Template C — Board pre-read

For the board of directors, 5–7 days before the meeting. Different audience, different norms, higher formality.
The Decisions needed from the board section is what distinguishes a board pre-read from a status update. Boards are decision-making bodies. If the pre-read does not surface decisions, the meeting becomes a download — a waste of the board.

Template D — Decision memo

For when a specific decision needs leadership input. Often ad-hoc.
The What I’m less sure about section is what distinguishes a good decision memo from a sales pitch. Leadership can tell when a memo is hiding the weak parts of an argument. Surfacing them honestly is what earns trust over time — and is the part that makes the memo most useful as a decision tool.

Save them as workspace templates

1

Start in the Composer with a real input set

For the weekly digest, do the manual input assembly described below, paste it into a new task, and iterate the prompt in the Composer until the draft lands.
2

Tune the constraints

Length caps, no-emoji, calibration discipline, the cap-at-three risk limit. The constraints are the template — they’re what produces the difference between a useful draft and a generic one.
3

Save with the right visibility

Weekly digest and decision memos are usually Private (the writer’s working surface). Business reviews and board pre-reads are often Workspace if multiple leaders will produce them with the same shape. See Custom templates.

Run a single readout

1

Assemble the inputs into a Customize context document

This is the manual synthesis step. For the weekly digest, gather:
  • The week’s key meeting artifacts (leadership meetings, all-hands, key project meetings, customer escalations) — surface them via the command menu (Cmd+K / Ctrl+K) and copy the relevant sections
  • Function-lead updates if you collect them
  • Metrics from your dashboards
  • Slack signal — what dominated the week, what surfaced as a crisis
Paste the inputs into a single Customize context document. This is the raw material the template runs on.
2

Write the headline first, before the draft

The headline is the artifact in compressed form. If you can’t write the headline, you don’t yet have a clear picture. A good test: if a leader read only the headline, would they have the right impression of the period? If yes, you’re done. If no, sharpen.
3

Run the template against the inputs

Draft from the saved template, with the assembled context document as input. This is the 80% draft — a starting point, not the final.
4

The 20-minute curation pass

This is where the writer’s judgment goes in. Five things:
  • Re-read the headline against the draft. Does the body support the headline? If not, either the headline is wrong or the body needs sharpening.
  • Calibrate confidence. Walk through every claim. “On track” claims that aren’t grounded should be downgraded to “tracking; the risk is X.” Hedge language (“we believe,” “should”) should be cut where the writer actually knows, and kept where they don’t.
  • Cut what isn’t useful. Hardest step. Every paragraph competes for attention. Anything that isn’t directly useful to the audience comes out, even if it took effort to produce.
  • Sharpen the asks. Vague asks (“input would be appreciated”) get vague non-responses. Specific asks (“approval by Friday on Option 2”) get decisions.
  • The leak test. Would you be comfortable if this readout reached someone outside the intended audience? If not, cut the sensitive content or restrict distribution.
5

Ship on schedule, archive in the durable home

Same day of the week, same time. Predictability is the commitment. After distribution, paste a copy into the series archive — Notion, Confluence, board portal, investor archive. The archive is what makes the series compound.
Triangulate signal in the synthesis. The highest-value content in a leadership artifact often lives in tensions: a function lead’s update says “on track” while customer call artifacts say customers are frustrated; the sales pipeline view says strong while recent deal artifacts say champions are weakening. Look for these tensions deliberately when assembling inputs. They’re invisible to single-source summaries; surfacing them is what makes the readout earn the read.

What this workflow doesn’t do

The biggest cross-meeting promise — “run a synthesis prompt across the week’s leadership meetings and get a structured digest” — is exactly the cross-call generative Q&A capability covered in the ad-hoc Q&A workflow. Single-call Q&A in the Composer is real today. Cross-call generative Q&A across a chosen corpus is not a one-click action. So the workaround above — manually assembling inputs into a Customize context document, then running the template prompt against it — is the actual workflow. The honest framing: the templates plus Composer save substantial time on the draft (the 80% step). The input assembly and the 20-minute curation pass are the writer’s work. Both together usually run 30–45 minutes for a weekly digest — meaningfully cheaper than producing the artifact from scratch. For larger-scale synthesis (quarterly reviews across many meetings), exporting local transcripts and running an external agent is the other option.

Distribution and feedback

Distribution rules by artifact:
  • Weekly digest — email or dedicated Slack channel to the leadership team, same time every week
  • Monthly business review — posted to the durable home; emailed to the exec team; often discussed at the next exec meeting
  • Board pre-read — through the board portal or directly to board members 5–7 days before the meeting
  • Investor update — email to the investor list, same day each month
  • Decision memo — directly to the decision-maker(s); followed up if no response within the requested timeframe
Watch for these signals to know whether the workflow is working:
  • Are leadership decisions being made faster? If asks turn into decisions within the cadence, the workflow is working.
  • Are board meetings less surprising? If pre-reads predict what comes up at the meeting, they’re doing their job.
  • Are leaders quoting the readout back? “As your weekly note mentioned…” is the highest-value feedback signal.
If none of these are happening after a quarter, something needs adjusting — the audience is wrong, the asks aren’t sharp, the cadence is off, or the content is missing what leadership actually needs.

The accuracy retrospective

Quarterly, the writer should look back at the previous quarter’s readouts and ask:
  • Where did the readout predict accurately?
  • Where did the read turn out wrong, and why?
  • Were there things we should have flagged but didn’t?
  • Were there things we flagged that turned out to be noise?
This retrospective sharpens the writer’s calibration over time. It’s also useful to share back with the audience: “Here’s what we said three months ago; here’s what actually happened; here’s what we’re updating.” The retrospective is manual today — pull the relevant past readouts from the archive, pair them with what actually happened, and run an external synthesis or work through them by hand. Cheap, high-leverage, and rarely done.

Common pitfalls

  • Writing for the writer, not the reader. The biggest single failure. The 20-minute curation pass is the antidote.
  • Bloated artifacts. Length is not value. Past a certain point, every additional paragraph degrades the signal of every previous paragraph. Hold the word counts.
  • Hedging instead of calibrating. “We believe we are tracking well” is hedge. “We expect to land at 110% of plan; the main risk is X” is calibration. Leadership notices.
  • Burying misses. Counterintuitively, surfacing misses with specifics builds credibility faster than reporting only wins.
  • Vague asks. “Your thoughts would be welcome” produces nothing. “Approval by Friday on Option 2” produces a decision.
  • Inconsistent cadence. A weekly digest that lands every Friday for 50 weeks beats a great digest that lands sometimes. Pick a sustainable cadence and hold it.
  • Trying to cover everything. The scope rule exists for a reason. Everything in the artifact crowds out the most important thing. Cut ruthlessly.
  • Letting the model write the final. The model produces the draft. The writer’s judgment — what to elevate, what to cut, how to calibrate — is the 20% that produces 80% of the value.
  • No durable home. Without the archive, the series doesn’t compound. Set up the home before the first artifact ships.
  • Confusing audiences. A weekly leadership digest written for the board reads wrong. A board pre-read written for investors reads wrong. Different audiences, different artifacts.
  • No retrospective. Without periodic review, the writer’s calibration doesn’t improve.
  • Leaks. Leadership artifacts contain sensitive content. The leak test before every send is non-negotiable. One badly-handled leak is worse than 100 missed cadences.

Where to go next