No app. No new number. One dial code.
Call G runs on conditional call forwarding— the same carrier feature that has powered voicemail since the 90s. Calls you don't answer divert to your personal Call G number, where your AI assistant picks up in under two seconds.
The call flow, end to end
- 1
Unknown number calls you
Your phone stays silent (or rings briefly, your choice). Saved contacts ring through normally — they never divert.
- 2
The call rolls to Call G
Carrier forwarding sends it to your local Call G number within about one ring. The assistant answers in under two seconds.
- 3
The AI screens, live
It discloses it's an AI, asks who's calling and why, and judges intent in natural conversation — not from a menu.
- 4
You get the right outcome
Real people: bridged to you with a whisper intro. Scams: blocked, with a transcript. Unsure: a structured message you can review.
What callers actually hear
Honesty is the policy — and the deterrent. The assistant opens every screening with full disclosure, in the manner you choose for it (friendly, formal, or brief):
“Hi — this is Margaret's AI assistant. Margaret doesn't take calls from unknown numbers directly. This call is recorded and analyzed. May I ask who's calling and what it's about?”
Genuine callers answer naturally and get connected or leave a message. Scammers face a calm conversational partner that asks the questions they can't answer — and hangs up the moment they ask for codes or payments.
Setting it up — both platforms, two minutes
iPhone
no app needed- Dial the forwarding code we give you, e.g.
*61*your-callg-number# - Turn on Settings → Phone → Silence Unknown Callers
- Unknown calls now roll silently to your assistant within one ring
Android
no app needed- Dial the same forwarding code — it works on every Android phone
- Optionally enable your dialler's “silence unknown callers” setting for zero rings
- Done — the onboarding wizard verifies it with a real test call
The wizard shows the exact codes for your carrier, then places a verification call so you see it working before you rely on it. Turning it off is one code too: ##002#
The rules the AI can't break
The decision policy lives in tested code, not in a prompt. These four rules are absolute:
Never blocks when unsure
Blocking requires very high scam confidence. Anything ambiguous becomes a message you review — a wrong block is the one mistake the system is built never to make.
Claimed emergencies always get through
If a caller says it's an emergency, they are connected immediately with a caution whispered to you first. No score overrides this rule.
Always says it's an AI
The first sentence of every screening discloses the assistant is an AI and that the call is recorded. Legally required (EU AI Act Art. 50) — and scammers often hang up right there.
Your contacts never meet it
Saved contacts ring your phone like nothing changed. The assistant only ever answers numbers you don't know.