Trust & how RedCues works
RedCues is built on one idea: the AI should help you, and it should never be able to lie on your behalf. These aren’t promises in a prompt — most are rules our build checks, so work that breaks them doesn’t ship.
The AI can’t make things up
It writes only from facts you’ve attested in your Library. With no source for a claim, it refuses rather than guesses — no invented figures, products or customers. Case studies, ROI and win-loss stories stay yours; the model may never author a customer engagement.
Nothing goes out without you
Every piece — content, a video, an email, a campaign — lands as a draft you approve. There is no “autonomous send” toggle. The AI drafts; a human decides and sends.
Your data is isolated, and it isn’t training anyone’s model
Each workspace is walled off at the database level, so one customer’s data cannot reach another’s. Your transcripts, contacts and content are sent to AI providers only to serve your own request, under API terms that exclude using it to train their models.
Recording asks first
Call recording requires your consent before it starts, and shows the two-party-consent disclosure where US law requires it. Opt-outs are honoured on every send — a contact who unsubscribes is removed before anything goes out.
The live cues are grounded, not a black box
Live in-call cues come from a deterministic engine graded against real calls — built to say the right thing at the right moment, and to stay quiet when it should. It never invents a claim mid-call.
What we don’t claim (yet)
We do not hold SOC 2 or ISO certifications today, and we won’t imply we do. If a security review is part of your decision, email hello@egama.ai and we’ll tell you exactly where we are rather than more than we have.