Introduction

Most scam calls are not clever. The same numbers dial thousands of people in a day, and the same recording plays to each of them. PhoneShield keeps a shared block list so that once a number has shown its hand to one customer, or to the FTC, it is refused for everyone before anybody answers. This guide explains where the list comes from, how strict each source is, and what to do when it is wrong.

Three sources, one check

The check runs between the forwarded call arriving and PhoneShield picking up. It is a lookup, not a conversation, which is why it is free and instant. Each source has its own bar for entry.

Your own blocked list

Numbers you have blocked yourself, from Shield settings or from a call’s detail. Private to your account. The ex-supplier and the persistent salesperson live here; they are not scams, they are just not welcome.

Reason shown: your block list

The shared block list

Numbers that have shown their hand to any PhoneShield customer. A confident scam verdict refuses the number on that customer’s line at once, and on every line once it is confirmed: by a second customer’s line, a known scam script, or FTC complaints. Softer failures need three strikes. Customer reports go here too.

Reason shown: scam verdict, three strikes, or reported

FTC Do Not Call complaints

Numbers with repeated complaints in the FTC’s Do Not Call data over the last 30 days. PhoneShield reads the FTC’s daily file every hour and refuses a number once its complaints cross a threshold.

Reason shown: FTC complaints, with the count

Two confirmations, or three strikes

The shared list has two speeds, and the difference is deliberate. A confident scam verdict refuses a number on the line that heard it straight away, and on every line once something confirms it. A softer failure has to happen three times.

A scam verdict, confirmed

A confident scam verdict comes from a caller actually saying the thing: the expiring warranty, the arrest warrant, the account that has been compromised, the prize that needs a fee. The number is refused on that customer’s line from the next call on. It is refused on every PhoneShield line as soon as any one of these agrees: a second customer’s front desk or Nora reaches the same verdict, what the caller said matches a scam script already confirmed elsewhere, or the number has complaints in the FTC’s data.

Silence takes three

Some calls fail without saying anything wrong. A predictive dialer connects and nobody is on the line. A caller launches into a sales pitch instead of saying who they are. Each has an innocent explanation once, so each counts as one strike. Three strikes, and the number is treated as the robocaller it almost certainly is.

Why wait for a second confirmation? Real scam campaigns dial thousands of people an hour, so a genuine scam number reaches a second PhoneShield line within minutes and is refused everywhere almost as fast as before. A single wrong verdict, though, now stays on one line instead of reaching every customer. In either case the first call is Shielded, not blocked: the front desk declined it, and the reason is on the call.

Safeguards against a wrong verdict

Some real calls sound exactly like scams. A bank really does call to confirm a wire you sent; a pharmacy really does call about a prescription. Wrong verdicts are rare and the system keeps improving, but because they can happen, three safeguards apply:

  • A single verdict stays on one line. As above: the network refuses a number only once a second line, a known script or the FTC's data agrees.
  • Saving a number vouches for it. When any customer saves a number as a contact or favorite, gives it an “expecting a call” pass, or imports it from their phone's contacts, that is strong evidence the caller is real. The shared block on that number is lifted, and it needs three fresh strikes to be blocked again. A scam script learned only from that number is dropped.
  • “Allow for me” is always available. Any customer can let a number on the shared list through for their own line without changing anyone else's protection (below).

Every entry carries its reason

Nothing goes on any list without a reason, and the reason follows the number everywhere. On the Shield settings page, each shared-list entry shows why it is there: a scam verdict, three strikes, a report with the category the reporter chose, or an FTC complaint count. In your call history, a refused call appears as BLOCKED with the same reason, and a declined one as SHIELDED with the classifier’s reason.

That lets you judge a block instead of trusting it: a pharmacy with an FTC count reads differently from a warranty scam with a verdict. It also keeps the list honest. Every entry traces to a call that happened or a complaint that was filed.

Refused before answering, at no cost

A blocked number is refused before the call is answered. There is no greeting, no classification and no Nora minutes used. It appears in your call history as BLOCKED with the reason, and counts toward the Blocked total on the Analytics page for that day.

What the caller hears

The call is not connected. Robocallers move on; there is nothing for them to interact with.

What you see

A BLOCKED line in call history with the reason, and a per-day count in Analytics. Blocked calls never ring you and never send a notification to Telegram, email, Slack or a webhook.

The FTC data, and why it is a threshold

The Federal Trade Commission publishes a daily file of Do Not Call complaints: roughly 12,000 a day, each a consumer saying a number called them unwanted. The complaints are unverified; nobody checks that the call happened or that the number was the real caller. PhoneShield reads the file every hour and keeps a rolling 30-day count per number.

Because a single complaint proves nothing, PhoneShield does not block on one. The block applies only when a number’s 30-day count crosses a threshold. Below the threshold the count is still recorded and shown on the call’s detail, so you can see a number that is starting to draw complaints before it is refused.

Three worked examples

Three kinds of number that reach the list.

A warranty robocall

A recording opens with the extended warranty on your vehicle. The front desk asks who is calling and why; the recording keeps going. The classifier returns a confident scam verdict. For the customer who received it, the call is Shielded with the reason “scam: vehicle warranty”. The number goes on the shared list the same second, and the next customer it dials sees only a BLOCKED line with “shared list: scam verdict”. No further conversation, no cost to anyone.

A pharmacy’s toll-free number with many complaints

A pharmacy chain runs refill reminders from one toll-free number. Thousands of patients get those calls, some no longer want them, and a share of those file Do Not Call complaints. In a busy month the number crosses the threshold and is refused with the reason “FTC complaints” and the count. That is right for most customers, who never asked for the calls. For a customer who fills prescriptions there, the fix is “Allow for me”, or adding the pharmacy as a contact or favorite. The block stays for everyone else.

A spoofed neighbor-prefix number

A scam dialer sets its caller ID to your area code and the first three digits of your own number, so the call looks local. The front desk hears the pitch, the classifier returns a scam verdict, and the number is shared-blocked. Two things follow. The dialer will soon move to another number, so the block matters most in the hours right after. And the number may belong to a real neighbour whose caller ID was borrowed. If that neighbour ever calls you, the reason on the block tells you what happened, and “Allow for me” lets them through for you alone.

The “Allow for me” override

Shared lists can be wrong about a number that is fine for you. You can allow such a number for your account without touching what everyone else sees.

  1. Open Shield settings in your dashboard

    Scroll to the shared block list. Each entry shows the number and the reason it is there.

  2. Find the number

    Use the search, or open the BLOCKED entry in your call history and follow it to the list.

  3. Click “Allow for me”

    The number gets a contact tag on your account. Contacts bypass the shared list for you only; everyone else keeps the block, and the entry stays on the list with its reason.

Reporting a number and preparing an FTC complaint

The list learns from calls automatically, but you can teach it too, and take a scam call to the FTC in a couple of clicks.

Report a number

Use “Report a number” on Shield settings when a caller was a scam and the front desk did not catch it. Choose the reason. The number goes on the shared block list with that reason and is refused for every customer from then on.

Prepare FTC complaint

On any Shielded call, click Prepare FTC complaint. PhoneShield copies a draft in the FTC’s own categories, with the number, date, time and what was said, and opens donotcall.gov, where you paste it in. The FTC has no filing API, so the last step is yours.

If a legitimate caller is refused

It happens occasionally, most often with the FTC source. Here is how to fix it.

  • Read the reason first. Open the BLOCKED entry in call history. “Your block list” means you blocked it yourself. “FTC complaints” with a count means a busy dialer, not necessarily a scam. “Scam verdict” means a caller from that number said something that reads as one.
  • Allow it for yourself. On Shield settings, find the entry and click “Allow for me”. The next call from that number reaches the front desk, or rings straight through if you also make it a favorite.
  • Do not expect the shared entry to disappear. The block is right for other customers who did not want those calls. Your allow is per account by design.
  • Tell the caller to identify themselves. A caller who states a name and reason at the front desk is not declined. Silence and pitches earn strikes.