Drafting and scheduling social posts with AI

How the bot works

Nine steps. Collecting the material, the approval queue, the per-platform formatting and the scheduling are ordinary code and cost almost nothing to run; the model is only asked to turn something you already published into a post.

1. It drafts from what you have already published. Your own pages, the service descriptions, the announcement you approved last month, the photographs you own, the document somebody wrote for customers. A library you can point at, not the open web.

2. It writes no claim that is not already yours. If your site does not say you are the fastest, the post does not either. Prices, availability, hours, hiring and anything about a product come from a source you named — and if there is no source, the draft says less rather than more.

3. Every draft shows what it came from. One line under each post: the page, the document or the announcement behind it. An approver can check a sentence in a second instead of wondering where it came from.

4. A week arrives at once, in your own mix. The number of posts, the types and the order are yours — and they arrive as a batch, because deciding about a week of posts once is a different job from deciding about one post seven times.

5. The approval workflow is the product. Drafts land in a queue with their sources and anything flagged at the top. A named approver reads, edits, approves or sends one back, in one pass, on a phone. Nothing moves without that pass.

6. Each platform gets its own version. Length, line breaks, how a link is handled, how many tags, what the image needs to be — one set of facts, written to fit each place it goes rather than pasted into all of them.

7. It schedules only what was approved. The approved text, at the times you set, inside your quiet hours and around the dates you blacked out. An edit made in the queue is what goes out; the draft it replaced does not resurface.

8. It replies to nobody. A comment, a direct message and a review are a person's to answer. The bot does not answer them, does not like anything, does not follow anybody and does not delete anything — it drafts and it schedules.

9. It tells you what it would not draft. The weeks it came up short and why: nothing new to say, a claim with no source, a photo with no rights record, a subject on your never-post list. A thin week is reported rather than filled.

What you configure

It can be configured to run in any business. This is what that actually means — the list of things that are yours, not ours.

Your material

Which pages, documents, photographs and announcements are the library — and what is out of bounds, including anything not published yet.

Your voice

Sentence length, how formal, whether you use emoji at all, what you call the customer, and the phrases that read as somebody else's brand.

Your banned claims

The superlatives, comparisons, guarantees and health, legal or financial claims that never appear in a post, enforced in the pipeline rather than suggested in a prompt.

Your mix and cadence

How many posts a week, which types in what proportion, what gets repeated and after how long, and which platform leads.

Your platforms and their formats

Which accounts are in scope, the length and tag rules for each, how links are handled, and what happens when an image does not fit.

Your approval workflow

Who approves, whether a second person sees it, how long a draft waits, and what happens when nobody approves — the answer is that nothing posts.

Your quiet hours and blackout dates

When nothing goes out, the dates you want silence, and what happens to a scheduled post when a blackout is declared.

Your images and your rights

Which libraries it may draw from, what has a usage record, whose face may appear, and what needs a credit.

Your release switches

Which accounts may be scheduled from an approved queue, and which require a person to press send every time. Replies stay off in every configuration.

Transparent, and configurable by you

The dark band below is what the system will not do. This is what you can see and change — and it is how we build everything, not something invented for this page.

You see what it did, and why

Every draft carries the source it was written from, every flagged phrase says which rule it hit, and every scheduled post records who approved that text and when. A post can be traced back to the page it came from rather than taken on trust.

Versioned prompts you can tune yourself

The voice, the banned-claims list, the mix and the prompts behind turning a page into a post are yours to tune. A change is a reviewable diff, replayed against weeks you have already approved so you can see exactly what it would have written differently, and rolled back the same way.

A human in the loop wherever judgment actually matters

Collecting material, drafting, formatting per platform, queueing and scheduling are automatic. Deciding what your business says in public, approving a post and answering a human being are a person's — that boundary is a property of the system, not a promise about how carefully we'll use it.

Every run logged, along with what it cost

One row per draft and one per scheduled post: the source, the rules it was checked against, what was flagged, who approved it, what they changed, when it went out and to which account — plus token usage and cost for that run.

It can be configured to run in any business. The material and the rules are yours: your library, your voice, your banned claims, your cadence, your approver, your quiet hours. The posting wiring on this page is one configuration of it, not the limit of it.

Works with your tools

This works with the accounts and the scheduler you already use. We're not selling you a social media platform, and the drafts stay yours in a format you can keep without us.

• A social scheduler you already pay for
• Facebook or Instagram
• LinkedIn
• Google Business Profile
• A website or CMS you control
• Google Drive, SharePoint or Dropbox
• A folder of your own photographs
• A spreadsheet of planned posts
• Microsoft 365 or Google Workspace
• Your own announcements and documents

Works well for

• Restaurants — Specials, hours and events change weekly, and the person who knows them is on the line at seven.
• Real estate agencies — Every listing and open house is a post, and the facts are already in the listing.
• Gyms and fitness studios — Schedules, classes and challenges repeat on a cycle that nobody has time to write twice.
• Salons and spas — Services, openings and seasonal offers are already written down somewhere else.
• Nonprofits — Programs, events and thank-yous need a steady voice that one person is holding together.

How a rollout works

Nothing posts on day one. The first milestone is a week of drafts from your own material, sitting in a queue, so you can read what it would have said before anything reaches an account.

1. 01Discovery

   The quote call, then a working session. We go through the material that counts as your library, your voice, the claims that never appear, how many posts a week and of what type, which accounts are in scope, who approves, and when nothing may go out. You get the library list, the rules and a written scope.

2. 02Dry run

   A week drafted against your real material. Nothing is scheduled. You read the posts, their sources and the flag list — and the flag list is where the work is, because a claim with no source is invisible until somebody asks where it came from.

3. 03Assisted mode

   Live on one account. The bot drafts into the queue, your approver does one pass, and approved posts go out on your schedule. Who decides what the business says does not change.

4. 04Optional limited automation

   Optional, and narrower than it sounds. A short list of post types you name, drafted and queued automatically on your cadence, still unscheduled until somebody approves, with every hard stop in force and the kill switch. Plenty of teams stop at step 03, and that's a good place to stop.

What we don't claim

• No numbers we haven't earned. No reach, engagement, follower or impression figure, and no posts-per-hour claim. We have not run your accounts. What posting steadily does for you is in your own analytics, and it is yours.
• No customer claims on this page. No names, no case studies, no logos.
• This is not a strategy. It drafts and schedules from what you have already decided to say. Deciding what your business should be saying, to whom, and why is a person's work — and if nobody has done that, a full queue will not fix it.
• A queue nobody approves posts nothing. That is the design, and it is also the honest risk. The approval pass is one job a week for a named person, and if that person does not exist this build goes quiet instead of going rogue.
• Consistency is not the same as growth. Showing up steadily is worth something, and it is not the same thing as an audience. We make no claim about the second one, and a number measured on somebody else's accounts would not have been about yours.

Questions people ask

Can AI run our social media for us?

It can do the part that is writing and scheduling. An agent can draft a week of posts from material you have already published, show you the source behind each one, and schedule the versions a named person approved. What it does not do is decide what your business should be saying, answer anybody, or spend money - so it is not the same thing as somebody running your accounts.

Where does it get what it posts?

A library you point at: your own pages, your service descriptions, the documents written for customers, the announcements you approved, the photographs you own. Not the open web, not a competitor's page, and not an invented fact about you. Every draft names the source it came from, and a claim with no source is flagged rather than written.

Who approves a post?

A named person, in an approval workflow that is the point of the build rather than a feature of it. Drafts wait in a queue with their sources and flags on top, one pass approves or edits a week's worth, and only the approved text is scheduled. If nobody approves, nothing goes out.

Will it reply to comments and messages?

No, and it has no path to. A comment, a direct message and a review are a person's to answer - answering reviews is a separate job we have a separate page for, and nothing here likes, follows, deletes or responds to anything.

Will this grow our following?

We make no claim about that, here or in a quote. Reach and engagement depend on your work, your pictures, your offer and a platform that changes its mind - none of which is a sentence this writes. What you get is a steady queue of posts you approved, drawn from things you have already said.

Can it sound like us?

That is the configuration. Sentence length, how formal, whether you use emoji at all, what you call your customers, the words you never use. The dry run exists so you can read a week in your own voice before anything is live, and the voice is tuned against posts you already approved.

Which platforms does it post to?

The accounts you name. Each one gets its own version inside its own format - length, line breaks, how a link is handled, how many tags, what the image has to be - rather than one block of text pasted everywhere. If you already pay for a scheduler, it can go out through that instead.

What about images?

It uses the libraries you point it at, and a photograph needs a usage record before it appears in a draft. A recognizable person needs a release on file. It does not generate a photograph of your premises, your staff or your work, because an invented picture is a claim about your business that nobody can stand behind.

Is there a record of what went out and who approved it?

Yes, per draft and per scheduled post: the source it was written from, the rules it was checked against, anything flagged, who approved that exact text, what they changed, when it published and to which account - plus the token usage and cost for that run.