How does AI keep my calls safe?

On Hello Fraands, AI follows every live call as it happens. Its job is simple: to end a call fast if it turns unsafe. No person listens to your call.

It understands intention, not keywords

Many systems only look for bad words. Ours does not use word lists. It tries to understand what the person means.

Here are two examples.

Example 1. Anita tells a host, “A man at my office said very bad things to me.” She is describing a bad experience. This is fine. She is not flagged.
Example 2. A caller starts with friendly chat. Slowly he says, “You sound like a nice person. Just tell me the OTP on your phone, it is for a small check.” This is a slow build-up to an OTP scam. The AI understands the intention, and acts.

The words are not the point. The intention is.

The call ends automatically if the AI sees:

harassment or abuse
threats
sexual pressure
a scam, such as asking for an OTP, UPI PIN, card details, bank details or Aadhaar number
“verify your account” links
someone pretending to be from a bank, telecom, courier, police or government
pushing loans, investments, products or remedies

When the AI ends a call, the account is flagged.

The call ends right away.
The account is flagged.
Our trust and safety team reviews the flag. A strike is confirmed only after this review.
If the person is a spammer or scammer, their account and phone number are blocked across the whole platform.

AI can be wrong sometimes. If your call was ended and you think it was a mistake, go to grievance page. A person will review it.

If you talk about self-harm, that is never a strike. We only do a welfare follow-up.

Either person can hang up at any time. Press # to end the call, block the other person and report the incident. You do not need to explain.