Elements
What your published posts are teaching the app. Each element of a post (hook, closing, posting time and so on) gets an estimate from real engagement. Statistics only propose findings; an editor confirms them before they guide new drafts.
Observational learning, not causal proof. Human selection and context may explain the associations.
Choose what to look at
Learning is kept separately for each brand and platform.
Human preference is shown separately and never used as a reward. These findings cannot change brand rules, evidence requirements or human approval.
Choose a brand to view its learning and version bank. Create one in Draft Lab if none are listed.