The difference between a generic answer and one you can publish is almost never the model. It is how much the assistant knows about your business before it starts writing.
Give it your brand first
Paste your website link, or describe the business in three or four lines, and ask for a brand profile. What comes back is a short dossier — how the business sounds, what it stands for, who it is talking to, and a few taglines in that voice.
That dossier is the point. Everything you ask for afterwards is written against it, so you stop re-explaining your tone in every message. Add your logo and your brand colours and the visuals follow them too.
If you have no website yet, a plain description works. The assistant does not need a brand book.
Your Brand Profile goes through it in full: building one from a site or from three lines, adding your colours and logo, and correcting it as you go.
Ask for the work, not only for a tool
Tool recommendations are the obvious use, and often the wrong one. The assistant also produces the output itself, which is usually what you actually wanted:
- a week of posts with captions, planned rather than improvised
- ad copy for Meta and Google, several angles at once
- replies to reviews and incoming messages, in your voice
- images and banners to go with the copy, on a plan
Asking for the finished piece and then editing it is faster than asking which tool would produce it and then going to that tool.
Attach your own files
Upload files next to the message box takes your own material — product shots, a price list, an existing post you want matched in tone. A photo of the actual product beats a stock image of a similar one, and it is the one thing no model can invent for you.
On a plan the assistant can work on an uploaded photo directly: turn it into a post, or put it on a different background.
Images and banners
Image generation is the part of the assistant that a plan pays for. Each image takes roughly half a minute and arrives in the same conversation as its caption, which is the whole reason to generate it here rather than in a separate image app.
On the free tier the assistant still writes everything — copy, plans, replies — it just does not draw.
Images and Banners covers what to ask for, working from your own photograph, and what to do about text inside a picture.
Revise in plain words
You do not need a new prompt to fix a result. Say what is wrong in ordinary language — too long, too formal, wrong colour, make the second option punchier — and it redoes that piece in the same conversation, with your brand profile still in place.
Starting a fresh chat throws that context away, so keep one conversation per campaign rather than one per message.
Voice, models and the rest of the box
The microphone beside Send lets you talk instead of typing, which is faster for briefs than for edits. The selector under the message box shows which model is answering; paid plans open up the stronger ones.
The theme switch follows your system setting by default and can be pinned to light or dark.
Keeping your conversations
Without signing in, a conversation belongs to the browser it was started in — clear the site data and it is gone. Sign in to sync chats ties them to an account instead, so the same history is available on another device.
Nothing you share is used to train models; it is used to produce your content and stays in your account.
What you may do with the output
You keep the rights to what you create here — see the Terms of Use. The limit is the obvious one: material you did not own going in, such as someone else’s logo or photograph, does not become yours because it passed through an AI tool.
When the assistant is not the answer
For anything about the service itself — billing, a plan that will not activate, a bug — write to the team from the contact page rather than asking the assistant. It knows marketing tools; it does not have access to your account.
Read Getting Started first if you have not opened the assistant yet.