Launch day
One announcement, shaped differently per platform, queued across the day instead of posted five times in one minute.
Nobody here does this full time, and hiring for it is not the next thing you need. Write the week in one sitting from the AI you already have open, and let it go out on its own.
Plan next week for us. Two LinkedIn, one Instagram. We shipped the new billing page.
Drafted, scheduled and queued in one exchange. You look at it again on Friday, or you do not.
Not for lack of a tool. Most companies already pay for one.
You post three times in a good week, then nothing for a month because something caught fire. The account looks abandoned right when someone checks it before a sales call.
A week gets written in one sitting and publishes on its own. Quiet weeks stop being visible from outside the company.
Writing the post takes twenty minutes, but deciding what to post takes the whole afternoon, so it slides to next week again.
Your assistant already knows what you shipped and how the account sounds. You approve or redirect instead of starting at a blank box.
It reads like a company wrote it, because a different person wrote each one and none of them agreed on a voice.
The voice comes from what the account has already published, so the posts sound like the same company on Monday and on Friday.
LinkedIn, Instagram, X and the rest. This is the only part that looks like setting up a tool.
A shipped feature, a hire, a customer story. One sentence is enough context for a week of posts.
Read them together, change the one that is off, and let the rest go. Editing three drafts is faster than writing one.
Posts publish on their schedule. Nothing needs a person present at nine on a Tuesday.
Most of a company year is quiet. These are the weeks worth getting right.
One announcement, shaped differently per platform, queued across the day instead of posted five times in one minute.
The steady drip that makes a young company look alive: what shipped, what broke, what you learned. It compounds precisely because it is boring to maintain.
A role posted the same week it opens, from the account candidates actually check, without waiting for someone to find the time.
The thing you meant to post about that great call and never did, written while you still remember the details.
The monthly changelog or the weekly tip that nobody owns, running on a schedule rather than on someone remembering.
Before, during and after a conference, queued in advance so nobody is drafting posts from a booth.
Automation is only comfortable if the wrong thing cannot publish itself.
Drafting and publishing are separate steps. A post sits as a draft until a person schedules or publishes it.
Disapproving a scheduled post removes it from the queue and cancels its slot, right up until it goes out.
Character limits and media rules are checked per platform, so a post does not get truncated into nonsense.
Posts are drafted against the account's own voice profile, which is a much easier thing to hand to a new hire than a tone-of-voice PDF.
Yes, deliberately. A four-person startup and a twelve-person company have the same problem: social matters, nobody owns it, and it is nobody's main job. The launch-day and build-in-public material above is the part written for you.
You work in Claude, ChatGPT or whichever assistant you already use. Antwork connects to it, so the drafting happens in a window you already have open. The web app is there when you want a calendar view, not as the place you live.
Two connected accounts and twenty posts a month, no card. That is a real month of posting for most companies at this size, not a seven-day trial.
It cannot publish on its own. Drafting and publishing are separate steps, and a scheduled post can be pulled out of the queue any time before it goes.
Then you already have the easy half. The part that does not get solved by a calendar is deciding what to write and writing it, which is where the week actually goes.
Two accounts and twenty posts a month, free, no card. Enough to find out whether this survives a real week.