AI
AI Content Studio
Content types, briefing, review/approval, and where generated content can be used.
What it is
The AI Content Studio is the same conversational assistant covered in the AI Chat / Assistant doc, running under its own agent profile so it also keeps a composition — a structured content draft — alongside the chat. The conversation can hold several compositions at once, each with its own draft id; the assistant edits a composition through tools, and you edit the same one directly in the workspace, both against the one document the server holds, so the two can never quietly drift out of sync.
3
Content kinds — social, email, document
30+
Named design size presets
3
Publishable platforms
Content types
Briefing and drafting
A design is generated as an HTML template with named placeholders plus the copy that fills them, rather than a single flattened image — which is what lets the workspace offer field-by-field editing afterwards, and what lets a design that worked be reused with new copy dropped into the same slots. Rendering happens in the browser, on demand, from that template and copy — nothing is drawn server-side. Size is resolved from a named preset (an Instagram square, a LinkedIn banner, an email header, and dozens more across every major platform) or from explicit dimensions if no preset applies.
Text in the chat is not the post
The assistant is expected to make every actual change to a post or email through its drafting tools, never by only describing the change in its reply — chat text can’t be scheduled or published, only what lands in the composition can.Review and approval
Publishing opens a dedicated review step that takes a one-way snapshot of the composition — edits made there refine that copy for sending, but do not write back into the assistant’s working draft. Nothing is rendered to a final image or uploaded until you actually choose to post or schedule; only at that point is each design rasterized and uploaded, so drafts you never send don’t pile up in storage. Once uploaded, the resulting file reference is written back onto the composition, so scheduling — and re-sending the same unchanged image later — reuses that same file instead of creating another.
- 1
Per-platform validation
Each target platform enforces its own limits before anything sends — LinkedIn allows up to 3,000 characters of text, Instagram 2,200 (and requires at least one image or video, and no more than 30 hashtags), Facebook up to roughly 63,000. - 2
Per-target results, not all-or-nothing
A post can go to more than one connected account at once; each target succeeds or fails independently, so one platform rejecting a post does not stop it reaching the others. - 3
Scheduling mints a fresh link at send time
A scheduled post stores the file’s storage key, not a direct link to it — temporary links expire before a distant schedule fires, so the backend generates a fresh one right before sending.
Where generated content can be used
A social post publishes immediately or on a schedule to whichever connected LinkedIn, Instagram or Facebook accounts you choose. An email draft can be sent as a campaign, or explicitly promoted into the brand’s reusable email template library — a one-way action you trigger yourself; nothing is added to the template library automatically. Buttons placed in a design or email only become real, trackable calls-to-action once you actually save that post or template — a call-to-action referenced but never resolved fails the save outright rather than shipping a dead button.
Related
New to briefing content? Start with the AI Content Studio guide. For the conversation and streaming mechanics underneath it, see AI Chat / Assistant.