Will the posts sound like me, or like AI?
Like you. Before it writes a single post, it reads posts of yours that you liked, and you tell it which words you never use. Every correction is kept: if you once deleted "indulgent", it won’t write it again. A post sounds like AI when the software was given only a topic, which is why the agent writes from your photo and your sentence. More on this in our guide to writing content with AI that sounds like you.
Does it create images and videos too?
Yes. The starting point is the photos you take, including on your phone: it crops them to each network’s dimensions, and if you have a designed template in Canva, it uses that. When there’s no suitable photo, it can create an image or a short video. Everything it creates comes to you for approval before anything is published.
Does it reply to customers in comments and messages?
On networks that allow it, yes. What each network allows is checked on the intro call. At first it drafts a reply from your product list and shipping terms, and you approve. Later you can decide that certain answers, such as price or delivery time, go out on their own in wording you approved. A complaint comes to you, and it doesn’t answer it. When most of the messages are customer questions, the right fit is an AI customer service agent.
Will it publish anything without asking me?
No. Every post comes to you as it will look, with an "Approve" button, and isn’t published until you tap it. If you decide later that one kind of post is published on its own, for example "back in stock", message us and we’ll set that up. One message cancels it.
What happens if it writes something wrong?
In the test on your own posts nothing is published, and you see everything. After that, too, every post goes through you before it’s published. It takes price, weight and ingredients from your product list, and doesn’t guess. If a wrong detail is published anyway, message it and it corrects the post, and we check what caused the mistake so it doesn’t happen again.
Who sees my photos and the messages from customers?
The photos and messages are kept on a separate cloud computer used only by your business, and aren’t mixed with other businesses’ information. To write a post or a reply, some of them pass to the AI company the agent writes with. When there’s sensitive information, we choose a company and terms under which it’s used for nothing else. On your pages the agent has its own permission, and you can revoke it at any time.
How much does social media management with an AI agent cost?
You pay once for the setup, then a monthly fee for ongoing support. There is also a charge for the services the agent uses, such as the AI company that writes and creates the images. You get an exact price in writing before we start. What the price is made up of is covered in our article How much does an AI agent for business cost.
How long until it starts working?
A first, simple thing, such as a post for every new product on one network, can be running within three days of connecting, with every post still waiting for your approval. All the networks together, with images, a weekly plan and replies to comments and messages, take one to two months, including the test on your own posts.
