Scale the scoring, keep the human send.
Roles control who can change weights, who can send, and who can enable automation, so the score scales without losing the human check.
The problem
As a scoring system spreads across a team, the risk is that someone flips on automation nobody agreed to, or reweights the model unnoticed.
What it does
Scoped roles
Separate who tunes the model from who works the tasks.
Send permission
Automated sending is a permission, not a default.
Auditable
Changes to weights and automation are logged.
How it works
Assign
Give each person the role that fits their job.
Gate
Sensitive actions require the right permission.
Track
Model and automation changes are recorded.
The human-in-the-loop promise holds at scale, because turning it off is a deliberate, permissioned act.
Scale the scoring. Keep the human send.
The short version.
Can I stop reps from changing weights?
Yes, tuning the model is a separate role from working tasks.
Is automated sending locked down?
It requires an explicit permission, so it cannot be switched on by accident.
Are changes tracked?
Yes, weight and automation changes are logged for audit.
See who’s hot right now.
Early access is open for a small design-partner group.