AI-generated copy
We generate four kinds of copy: cold outbound drafts, weekly Monday report copy, rubric rationale text shown next to the score, and on-page example posts on /channels. Every one of them is drafted against your active ICP rubric — industry, headcount band, role keywords, spend floor — not against a randomised prompt.
What we do NOT generate: subject lines from a randomised prompt. Subject lines are pulled from the prospect context the agent already has (their industry and the rubric criterion they matched), or written by the human reviewer.
Team and Scale accounts route every cold-outbound draft through the human-in-the-loop review queue before it goes out. The Monday report names how many drafts were approved and how many were auto-sent after their HITL window expired. Solo accounts opt into the same queue under /settings/hitl.