Verified against ChatGPT · 2026-08-14
Draft an internal change announcement that survives the reply-all pile-on
Writes an internal announcement of an operational change with the likely objections pre-empted in the text itself, so it doesn't spiral into a reply-all thread of confusion and complaints.
The prompt
Ready to copy — highlighted parts are example details you can swap.
Write an internal announcement for an operational change, structured to pre-empt the confusion and complaints it will predictably generate rather than triggering a reply-all thread.
WHAT'S CHANGING
Switching the expense reimbursement process from monthly batch approval to a rolling weekly approval cycle.
WHO IS AFFECTED
All employees who submit expense reports, roughly 400 people across three departments.
WHY THIS IS CHANGING
The monthly batch process was causing a two-week backlog that delayed reimbursements past the point employees were fronting real money.
EFFECTIVE DATE
Starting the first Monday of next month.
WHO TO CONTACT WITH QUESTIONS
The Finance Ops team via the #finance-help Slack channel, not individual finance staff.
Open with what's changing and the effective date in the first two sentences — do not bury the actual news under throat-clearing about company values or a long preamble; people affected by a change scan for 'what changed and when' first and read the reasoning only if that's already clear. State the reason for the change honestly and specifically enough to be believed — a reason that sounds like corporate boilerplate ('to better serve our stakeholders') invites more suspicion and pushback than a plain, specific one, even when the plain one is less flattering. Anticipate the two or three questions this specific group of affected people will actually have — not generic FAQ filler — and answer them directly in the body, since an announcement that leaves an obvious question unanswered is what actually triggers a reply-all thread, as people ask it publicly instead of privately. Name the single point of contact clearly and make it the only channel offered, so questions route to one place instead of scattering across replies, DMs, and hallway conversations that create conflicting answers.
WHAT NOT TO DO
Do not use passive voice to soften who made the decision ('it has been decided that...') — state who owns the decision. Do not promise reassurance you can't back up ('this won't affect your workload') unless that's actually true; an announcement caught overpromising loses credibility for every future one.
OUTPUT FORMAT
A ready-to-send announcement: subject line, opening two sentences stating the change and date, a short reasoning paragraph, a 2-3 item anticipated-question section answered directly, and a closing line naming the single contact point.Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
Reply-all pile-ons on internal announcements are, mechanically, almost always caused by an obvious question left unanswered in the original message — someone asks it publicly because there was no other visible channel, and once one person replies-all, others pile on with their own questions in the same thread; instructing the model to specifically anticipate this group's likely questions and answer them inline, rather than writing a generic body paragraph, closes off the exact trigger condition. GPT-5.1's default register for internal announcements leans toward soft, values-forward corporate phrasing ('to better serve our team') when not given a concrete reason, and that vagueness reads as evasive to an audience already primed to be annoyed by a process change — requiring the reasoning to be specific and honest, even when it's less flattering, produces language that reads as trustworthy specifically because it doesn't sound like boilerplate. The instruction to open with the change and date in the first two sentences targets how people actually read mass internal emails: they scan for personal relevance before reading justification, so an announcement that leads with preamble gets skimmed past its own reasoning entirely and the reader goes straight to asking the question the announcement was trying to pre-empt. Naming exactly one contact channel, and instructing the model not to let multiple people appear as valid contacts, prevents the scattered-and-conflicting-answers problem that happens when questions get fielded by whoever happens to reply first, which is often not the person with the actual correct answer.
What you get back
Subject: Expense approvals move to weekly starting [date]. Starting the first Monday of next month, expense reimbursements will be approved weekly instead of in a monthly batch. This is changing because the monthly cycle was creating a two-week backlog that left people fronting real money longer than it should. If you're mid-cycle when this switches, your current report will be processed under the new weekly schedule automatically. Questions go to #finance-help, not individual Finance staff.
Verified against
ChatGPT GPT-5.1 · 2026-08-14
Changelog
- 2026-08-14 — Initial publish, verified against ChatGPT GPT-5.1.
Need this built into your business?
If a prompt isn't enough — custom software, built and maintained for you — that's Scult's day job.
EXPLORE CUSTOM SOFTWARE
