Internal Comms & Meetings

How to draft a slack announcement with AI

Internal Comms & Meetings3 AI tools6 steps6 friction points

Want an agent to handle this on your live data? See the Starch version →

Drafting a Slack announcement sits at the intersection of writing and judgment. You need to say the right thing, to the right people, in the right tone — whether you're announcing a product launch, a policy change, an org update, or a win worth celebrating. Most operators write these from scratch each time, second-guessing word choice and structure while also trying to remember what actually happened and who needs to know.

The reason people reach for AI here is obvious: the hard part isn't writing, it's writing something that sounds right for your audience and doesn't read like a press release. AI is good at tone adjustment, at compressing a long context into a short announcement, and at generating three different versions when you're not sure which angle to lead with. The skeleton of a good Slack message is a pattern AI can hold.

ChatGPT, Claude, and Gemini can all draft a solid Slack announcement if you give them enough context. Paste in your meeting notes, a bullet list of what happened, or a rough draft — and any of these tools will return something usable in one or two exchanges. The output is often cleaner than what you'd write under time pressure. The real limitations show up in what you have to do before and after the prompt.

Skip the manual work → See the Starch version

Internal Comms & Meetings3 AI tools6 steps6 friction points

AI walkthrough

How to do it with AI today

A practical walkthrough using ChatGPT, Claude, and other off-the-shelf LLMs — what they're good at, what you'll have to do by hand.

Tools that work for this

ChatGPTClaudeGemini

Step-by-step

1 Gather your raw material before touching any AI tool — this is where most people underinvest. Pull together the relevant context: the decision that was made, who it affects, the timeline, and any action items for the audience. A bullet list works fine.

2 Open ChatGPT, Claude, or Gemini and paste your context in along with a prompt that specifies the channel, the audience, and the tone you want. Channel and audience matter more than most prompt guides acknowledge — an all-hands announcement reads differently than a message to #engineering.

3 Review the first draft for accuracy, not just style. LLMs will produce fluent prose that may subtly misrepresent what you said. Read it against your original notes and correct anything that drifted.

4 If you need multiple versions — a short version for #general and a detailed one for the relevant team channel — ask for both in the same session. Describe the length and framing difference explicitly; don't assume the model infers it.

5 Copy the final draft into Slack manually, double-check any @mentions or channel references the model included (it can't know who's actually in your workspace), and send. There is no direct posting path — you're always the last link in the chain.

6 If this announcement will recur — weekly team updates, sprint summaries, monthly metrics — save your prompt and your context-gathering checklist somewhere. Next time you'll be reconstructing both from scratch unless you do.

Prompts you can copy

Draft a Slack announcement for #general that our mobile app v2.3 ships Friday. Audience is the whole company, ~30 people. Tone: excited but not hypey, two short paragraphs, end with one action item for people to test.

We're changing our PTO policy starting next quarter: employees now get 20 days instead of 15. Write a Slack message for #announcements that explains what changed, why, and what employees need to do (nothing — it's automatic). Keep it under 100 words.

Summarize these meeting notes into a Slack update for the #product channel. Focus on decisions made and open questions. Audience is 8 engineers and 2 PMs. [paste notes]

We hit $1M ARR today. Write a Slack message for #wins that acknowledges the team, names three specific people who drove it, and sets up what comes next. Don't make it sound like a LinkedIn post.

Rewrite this draft Slack announcement in a calmer, more direct tone. Remove the exclamation points. The audience is skeptical ops folks who just want the facts. [paste draft]

Reality check

Where this gets hard

The walkthrough above works — until your numbers change, the LLM hallucinates, or you have to re-paste everything next month.

No connection to your actual context — you manually copy in meeting notes, metrics, or decisions every single time, and anything you forget to paste, the model doesn't know.

The model has no memory of your team, your channels, your recurring announcements, or your preferred formats — you re-explain the same organizational context in every session.

Outputs vary run to run. The structure and tone you carefully calibrated last sprint aren't guaranteed to reappear when you run the same prompt three weeks later.

Nothing persists downstream. After you get the draft, you're back to Slack manually — no scheduling, no sending, no logging that the announcement went out or who saw it.

For announcements that depend on live data — last week's metrics, a sprint velocity number, a revenue milestone — you have to pull those numbers yourself before the prompt. The model can't reach into Stripe or your project tracker.

One-off workflows don't compound. If you draft team updates every Monday, you're rebuilding the same prompt context from scratch each week rather than running a process that gets smarter over time.

Tired of the friction?

Starch runs the whole workflow on live data — no copy-paste, no hallucinated numbers, no re-prompting next month.

See the Starch version →

Starch alternative

The same workflow on Starch

Starch is an agentic operating system — for this workflow, that means an agent builds a persistent app that pulls live context from your actual business data, drafts the announcement, and posts it to Slack on a schedule, without you reassembling the prompt chain each time.

Connect Slack as a scheduled-sync provider once — Starch knows your real channel names and workspace structure, so the agent drafts announcements targeting the right audiences without you supplying that context every run.

Pull live context automatically before drafting. Connect Google Calendar, Notion, or Gmail from Starch's integration catalog, and the agent reads what actually happened this week — meetings, decisions, deliverables — before it writes a word.

Use the Meeting Notes app to capture decisions and action items in real time, then tell Starch: 'Every Friday at 4pm, take this week's meeting summaries and draft a #general Slack update with decisions made and next steps.' Starch builds that automation and runs it on schedule.

The Knowledge Management app keeps your company context — team structure, recurring update formats, product terminology — in one searchable place the agent reads before drafting, so announcements don't drift in tone or miss key context.

Describe the full workflow in plain English and Starch builds the persistent app: 'Each Monday, pull last week's metrics from Stripe, summarize them in three bullets, and draft a #wins Slack message for my approval before sending.' That runs continuously — you review and approve, not reconstruct.

Get closed-beta access →

Toolkit