RedHub AI
How the gate works
How the Brand-Voice gate works
This tool grades an AI-drafted piece against your own brand voice. Each signal is scored 0–2; the verdict is the weakest one — then a gate that can only make it worse.
The 6 signals
A draft that reads like a machine — generic AI-slop phrasing, hollow cadence — is demoted one band, even when every dimension marks on-voice: ON-VOICE reads DRIFTING and DRIFTING reads OFF-BRAND. A draft already OFF-BRAND stays OFF-BRAND. The flag only ever lowers the read.
Why: The sample draft, Email — nurture #2, marks on-voice (2) on every one of the six voice dimensions, so its weakest dimension reads ON-VOICE. But the draft carries the anti-slop flag: it reads like a machine. The gate demotes it one band, so the read is DRIFTING.
A clean sheet didn’t save it. The verdict is the weakest signal — and a top trigger is dispositive on its own — so where a signal is short, the fix first row names it. Grades a draft, never a person. Runs offline in the engine — nothing is fetched, sent, or published.