ع
Learn Tracks Reference Guides Saved
playbook

The weekly stakeholder & investor update system

Turn a metrics export, a few Slack threads, and your own brain-dump into a crisp, on-message update — internal or investor — in about 20 minutes, and faster every week.

easy ~20 min, weekly
when to reach for this

The recurring update is the chore every founder and PM dreads and quietly skips — and skipping it is how investors go cold and a team loses the plot. The fix isn't discipline, it's a system: a fixed template plus your product context, so each week is a 20-minute fill-in, not a blank-page slog. Feed in the raw week — a metrics export, the Slack threads, what's in your head — and Claude drafts a clean update against that template. Because the template and context are reused, it compounds: week two is faster than week one, and the same raw week can become a candid internal update or a concise investor update without redoing the work. It's all a conversation in Claude Desktop — no terminal, no install.

gather this first
  • Your update-template.md — built once in step 1, reused every week after.
  • This week's raw inputs: a metrics export (metrics.csv or a pasted table), the Slack/standup threads that mattered, and a quick brain-dump of what happened — wins, fires, decisions. Drop the files into the folder you open in Claude Desktop, or paste the lot into the chat.
  • A product-context.md if you have one (what you're building, for whom, the metrics that matter, the stage you're at) — it keeps the draft on-message instead of generic.
  • Last week's sent update, so Claude can track the deltas and pick up where you left off.
the workflow

  1. Build the template once

    Define the shape of every future update a single time. A fixed skeleton is what turns a dreaded blank page into a fill-in-the-blanks task — and it's the thing that makes the rest of this fast. Have Claude draft it, then you decide the final sections.

    you ask
    Draft a reusable weekly update template as update-template.md with these sections, each with a one-line note on what goes in it: (1) Metrics block — the 4-6 numbers we track, with last-week and target columns; (2) Wins — what moved; (3) Lowlights & risks — what's stuck or worrying, stated plainly; (4) Asks — where I need help, intros, or decisions; (5) What's next — the focus for the coming week. Keep it tight and skimmable.

    what you get back A clean update-template.md skeleton with the five sections and guidance notes — the fixed frame you'll pour each week into.

    Save this file. Tweak the metrics block to the 4-6 numbers you actually live by; this template is the asset that makes every future week 20 minutes instead of an hour.

  2. Draft the update from the week's raw inputs

    Now feed Claude the messy week plus the template and your product context, and let it do the assembly. You're not writing — you're handing over the raw material and getting a structured first draft back to react to.

    you ask
    Using update-template.md and product-context.md, draft this week's update. Here are the raw inputs: [paste the metrics export, the key Slack threads, and my brain-dump of what happened]. Fill every section. For the metrics block, only use numbers I actually gave you — if a number is missing, leave a [TODO: confirm] placeholder rather than inventing one. Flag anything you're inferring.

    what you get back A complete first draft mapped onto your template — metrics filled from your data, wins and lowlights pulled from the raw inputs, and honest [TODO: confirm] gaps where a number wasn't supplied instead of a confident guess.

    The placeholder instruction matters: it's the difference between a gap you'll notice and fill, and a fabricated figure you'll send by accident.

  3. Tighten for the audience and the tone

    This is the move that earns its keep: the same draft becomes two different updates. An internal update is candid, detailed, and action-focused; an investor update is concise, metrics-first, and leads with the headline. Tell Claude which one you're shaping and have it cut accordingly.

    you ask
    Tighten this into an INVESTOR update: lead with the 3-4 headline metrics and the one-line story they tell, keep wins to the few that matter, state lowlights and risks honestly (don't spin them), and make the asks specific and easy to act on. Cut anything that reads as internal noise or filler. Then give me an INTERNAL version of the same week: more detail, the operational lowlights, and the team-level next steps. Both under one screen.

    what you get back Two tuned versions of one week — a concise, metrics-first investor update and a fuller, candid internal one — with the fluff cut and the asks made concrete in each.

    Read both yourself before anything ships. Claude tightens prose well; it can't know which lowlight you're not ready to put in front of the board this week. That call is yours.

  4. Track the deltas and save the sent version

    What makes this a system and not just a weekly draft is memory. Have Claude compare against last week so the trend is visible — "MRR up 8%, churn flat, two asks from last week still open" — then save what you actually sent, so next week starts from it instead of from scratch.

    you ask
    Compare this week's update to last week's [paste it]: which metrics moved and by how much, which asks from last week are still open, and which lowlights resolved or got worse. Add a one-line 'since last week' delta to the top of the update. Then save the final version I'm sending as update-2026-06-24.md.

    what you get back A delta line that makes the trajectory obvious at a glance, a check on whether last week's asks got actioned, and the sent update saved as a dated file — the seed for next week's draft.

    Saving the sent version (not the draft) is what lets the loop compound. Next week, last week's sent file is your starting point and your accountability check.

make it your own
  • Adapt the cadence to the audience. A team update can run weekly and detailed; an investor update often lands better monthly and tighter. Build the template once, then run the investor cut every fourth week from the four internal updates Claude already has — it summarizes the month instead of you re-gathering it.
  • Write the Arabic update natively, don't translate it. If your board, family office, or local investors read Arabic, don't machine-translate the English. Draft that version directly in Arabic from the same raw inputs — Arabic carries its own register and formality for stakeholder communication, and a translated English update reads stiff and off-key. Same numbers, same honesty, written for the reader.
  • Investor-relations memory. Keep a running investors.md of who's on the list, what each cares about, and the asks you've made — so Claude can tailor the asks section and remind you what you promised whom.
  • Automate the assembly (Power Track): once the prompt chain is dialed in, advanced teams can save it as a /update custom command or wire a scheduled agent to draft the skeleton every Friday (see the Features tab). On Desktop you can keep the dialed-in prompt in a note and paste it each week instead.
watch out for
  • Verify every metric and number before it ships. Claude states figures with total confidence, including ones it inferred or carried over wrong. An investor update with a wrong number erodes trust fast and is hard to walk back — check each figure against the source export yourself. The [TODO: confirm] habit from step 2 is your safety net; don't paper over the gaps, fill them.
  • Never paste confidential financials, cap-table details, or customer PII into a general chat. Keep the sensitive raw data — full P&L, the cap table, named customer revenue — inside a workspace your company has approved for it, and share only the aggregates (MRR, growth rate, runway in months) with the chat. The update should carry the headline number, not the underlying confidential file.
  • Report lowlights honestly — don't let the draft spin. Claude will happily soften a bad week into something rosier than it was. A polished update that hides the churn spike or the slipped launch is worse than a blunt one: investors and teams forgive problems, not surprises. Make sure the lowlights section says what actually went wrong.
  • The update is yours to own and send. Claude drafts; you decide what's in it, how candid the lowlights are, and which asks to make. Read the whole thing before it goes out — your name is on it, not Claude's.

you'll end up with A repeatable update engine: a raw, scattered week becomes a sent, on-message update — internal or investor — in about 20 minutes, consistent every cycle, and faster as the template and your sent history compound week over week.

Questions people ask

How does this make writing the update faster instead of just moving the work around?
The speed comes from reuse. You build the template and your product context once, and after that each week is Claude filling a fixed skeleton from your raw inputs rather than you facing a blank page. By week three you're reviewing and verifying a draft, not writing one — and last week's sent update seeds this week's, so the gathering shrinks too.
Can one draft really serve both my team and my investors?
Yes — that's the core move in step 3. The same raw week gets tightened two ways: an internal update that's candid, detailed, and action-focused, and an investor update that's concise, metrics-first, and leads with the headline. You shape both from one draft instead of writing two from scratch.
Is it safe to put my financials and metrics into Claude for this?
Share aggregates, not confidential raw data. Headline metrics like MRR, growth rate, and runway are fine to draft with. Your full P&L, cap table, and named-customer revenue should stay in a workspace your company has approved for that data — paste the summary number into the chat, never the underlying confidential file or customer PII.
Will Claude get my numbers right?
Treat every figure as unverified until you've checked it against the source. Claude states numbers confidently even when it inferred or mis-carried one, and a wrong figure in an investor update is costly to walk back. This playbook has Claude leave a `[TODO: confirm]` placeholder for any number you didn't supply — your job is to fill and verify each one before sending.