AI manager and content breakdown
- The AI manager: what it is and how it differs from “rewrite”
- A content plan from the model
- Content breakdown: why a post worked
- Competitor topics
- Adapting a post for each channel
The AI manager: what it is and how it differs from “rewrite”
The “AI manager” section hands the model the whole network rather than a single text: channels, posts, subscribers, channel health, money, competitors and the empty days in the plan. That is why it answers “what happened”, “why the numbers dropped”, “what to post about”, “how the money is doing” and “client report”. Next to the answer there is a “What the model saw” button holding exactly the numbers it was given, so any figure in the answer can be checked. The manager cannot publish: everything it creates is a draft in the editor.
A content plan from the model
The “Posting plan” tab: the dashboard itself lays the slots out across days and channels — from your posting rhythm over the past month and each channel’s best hours — and the model suggests a topic for every slot, alternating formats (news, how-to, engagement, call to action). Tick the rows you like and press “Create drafts”: up to ten posts at a time, each written in the channel’s own voice from the AI memory. A draft goes nowhere until you queue it yourself.
Content breakdown: why a post worked
The “Content breakdown” block on “Analytics” compares your own posts with each other: length, a hook in the first line, media, a call to action, a question, a list, emoji, hashtags and the hour of publication — and shows how much better posts with that trait do than the rest in your channels. It is computed on views (a userbot is required), otherwise on joins within 24 hours. Fewer than 12 posts for the period and there are no conclusions: on five posts any “finding” is chance. The “Copy the style for AI” button gives a description of the winning style you can paste into the notes for the model.
Competitor topics
The “What they post about” block in “Competitors” computes topics from the posts the userbot has already read: a topic makes the list once the same thing has been written about at least three times and in more than one channel. Next to it, “read ×N” is the average views of the topic’s posts against the average across every post read. An observation, not a forecast. Topics also go into the daily competitor digest in the bot.
Adapting a post for each channel
In the editor, with two or more channels selected, the “Per-channel edits” block has an “Adapt with AI” button: the model rewrites the text separately for each selected channel, leaning on that channel’s memory — its style profile and samples of its best posts. If the post has media, the variant is shortened to the 1024-character caption limit. A failure on one channel does not affect the others.
two channels free, no card required