Claude for
leadership work
Board prep, strategy memos, all-hands drafts, and the synthesis work that sits between a stack of inputs and a decision.
Leadership work is a constant conversion problem: take the inputs from the executive team, the board, the market, and the team, and turn them into a small number of clear decisions. Claude cannot make the decisions. It can do almost all of the conversion. The habits on this page are about using it as a thinking partner without accidentally handing it your judgement.
Three places to start
Three workflows where Claude does the synthesis work without touching the call.
1. Drafting a strategy memo from rough bullets
Your thinking usually starts as rough bullets from a walk or a long meeting. Claude will turn those into a proper memo that tracks what you actually said. The critical instructions are "do not add anything I did not say" and "where my thinking is incomplete, flag it as an open question rather than filling in plausibly". The open questions are the agenda for your next exec meeting.
Strategy memo prompt, annotated
Four elements, pointed specifically at the "don't drift from what I meant" problem.
Where this falls over
Without the open-question instruction, Claude will paper over the gaps in your thinking by inventing plausible middle-ground positions. The memo looks coherent but drifts from what you actually meant.
2. Synthesising written input from the exec team
You asked the exec team for written responses to a strategic question. You got eight responses of varying length and quality. Claude will summarise each fairly, lay out where people agree and disagree, and identify the core tension. Crucially, ask it to end with your honest read of which way the argument is leaning and what questions would resolve it. The synthesis does the work of the meeting before you have the meeting.
Synthesising exec input, weak vs strong
The weak version produces a polite blur. The strong version produces the meeting you need to have.
A polite overview that smooths every disagreement into "there are different views on this, and a balanced approach might consider...". Every exec sounds reasonable and no one sounds wrong. Unusable as the input to a decision.
Where the team agrees: Europe is strategically important. Entry is a question of when and how, not whether.
Where the team disagrees: The split is not 4-4. It is 5 people who want to enter in Q2 with a direct team, 2 people who want to enter via partnership first, and 1 person (the CFO) who believes the cost case has not been made and wants a formal business case before committing.
The core tension: this is not "fast vs slow". It is "should we bet now on execution speed, or should we bet on reducing execution risk first". Anyone reading the responses can see it phrased as the former, but the CFO's response reframes it as the latter, and the rest of the team has not engaged with that reframing.
Questions that would resolve it: (1) what does a minimum defensible business case look like for the CFO to withdraw the objection? (2) if the answer is "it does not exist yet", who owns producing it, and by when? (3) what is the cost of waiting 6 weeks to let the CFO's concern be addressed versus committing now?
Where this falls over
Claude's default is to split the difference for politeness. The explicit instruction must be: do not split the difference, state the disagreement fairly.
3. All-hands drafts that people actually read
Monthly all-hands updates are a load-bearing document. The good ones are warm, specific, and honest about bad news. The bad ones get skimmed. Claude is surprisingly good at the good version if you give it examples of past updates you were happy with and tell it to lead with the one thing that most matters, not the one that's easiest to write.
First month of using Claude for the all-hands draft. Process: paste the month's wins, losses, and tricky bits into a Project with my last three all-hands messages as voice references. Ask for a draft with the explicit instruction "lead with the single most important thing, do not bury bad news, keep it under 500 words".
First draft tried to bury a hiring freeze at the bottom. I pushed back once and asked it to move that to the top paragraph. The second draft opened with the freeze, explained why, and set out what stays the same. That is now the message going out Monday. Less polished than if I had written it from scratch, more honest than my last three attempts.
Where this falls over
Left to its own devices, Claude will bury bad news in the middle and lead with the easy good news. Specifically tell it not to bury bad news, and push back on the first draft if it does.
Worked examples
Before and after, on tasks from leadership work.
Prompts to steal
Copy any of these, fill in the square brackets, run.
Try this week
Pick one. All of these are under an hour.
The boundary for leadership work is tighter than for any other role: Claude does the writing, never the thinking, and never the deciding. Within that boundary, it can save you a remarkable share of the hours you currently spend drafting. Use the saved time on the decisions themselves.
If you want to go broader, the Beginner guide covers the habits that apply everywhere. The prompt library has the full set of role-tagged prompts in one place.