Solution: Anti-vishing
Anti-vishing that works on the line, not in a training deck
Vishing is social engineering by phone call. Most advice about stopping it tells the person to hang up and call back. Open Voice Shield stops it on the line instead: every call in or out is recorded, transcribed and scored, a flagged number's next attempt is blocked, and the recording, transcript and report are kept as evidence. The verdict lands minutes after hangup, or during the call in realtime mode. It runs as one hop on the existing phone system or trunk, with nothing to install.
What vishing is
A call that talks a person past every other control
Voice phishing, vishing, is a phone call in which the caller pretends to be someone the person trusts and talks them into handing over something: a password, the one-time code that just arrived by text, card details, a payment to "new" supplier bank details, or remote access to a computer. The scripts are rehearsed, calm and urgent: the bank's fraud desk verifying a card, the IT help desk fixing an account, a government agency with a warrant, the owner travelling and needing a transfer before noon.
It works for two reasons. A call feels personal, and pressure works better in a conversation than in an e-mail. And the phone line has none of the filters e-mail has: no record of what was said, no score, no alert, often no memory beyond "someone from the bank called". Caller-ID authentication, required by the CRTC since 2021 and described in the Canadian Centre for Cyber Security's guidance on vishing, says whether the number was forged. It never hears the call. Read the scripts scammers use and how vishing differs from phishing and smishing.
For everyone else: what vishing is, how to prevent it at home, at work and on a network, and what to do after a vishing call if something was already shared.
Why the usual advice is not a defence
"Hang up and call back" depends on the person
The advice is right, but it is not a control. It assumes the person notices, in the moment, under a deadline, with a confident voice on the line, that this is the call to hang up on. Some do. The scammer needs one who does not. Awareness training and simulations raise the odds. They do not record the call, alert anyone when the real one comes, or stop the second attempt an hour later to a different desk. A business that has locked down e-mail, the website and the database has usually left the phone line exactly as it was. The gap is laid out here.
| Measure | Judges | Stops the first call | Stops the next attempt | Keeps evidence |
|---|---|---|---|---|
| Awareness training, simulations | the person's reaction | sometimes | no | no |
| Caller-ID apps, "scam likely" labels | the number's reputation | known numbers only | known numbers only | no |
| STIR/SHAKEN attestation | whether the number was forged | forged numbers only | no | no |
| Open Voice Shield | the conversation itself | in realtime mode, on request; standard mode scores after hangup | yes, automatically | recording, transcript, report |
How it works
Every call scored, the next attempt stopped, the evidence kept
- One hop on the line you have. A business's phone system or SIP trunk sends calls through Open Voice Shield on the way to the carrier. A carrier routes a client, a prefix or a sample of traffic through it. Phones, numbers and call flows stay as they are. For a business, for an operator, for a mobile operator.
- A verdict within minutes. Each call is recorded with the caller and callee on separate channels, transcribed, and scored: a probability from 0 to 100, the category (bank or card fraud, government or IT impersonation, supplier payment redirection, pressure sales, robocall), the entity the caller claimed to be, numbered red flags quoting what was said, and block, review or allow. A block needs a hard sign in the transcript: a request for money, cards or codes, a threat, a deadline, impersonation or remote access. How the detection works.
- An alert where the team already is. E-mail, Slack or a signed webhook into ticketing or a SIEM, above the threshold you set.
- The next attempt blocked. An automation rule puts a flagged number under control: its next call is refused or diverted to voicemail, a security line or an AI agent, for an hour, a week or for good. Most voice scams take more than one call.
- Evidence that holds up. The stereo recording, the timed transcript, the SIP log, a PDF report, and a password-protected share link for the bank, the insurer or the police.
Limits
Limits of the verdict
Standard mode after the call, realtime mode during it
In standard mode the verdict arrives minutes after hangup. What it stops is the next attempt from that number, and what it gives the person is a record of what was said. Realtime mode scores the call while it is still up, in time to act before it ends. It is set up on request and costs more, because the whole call is transcribed and read as it happens.
Either way, the probability is a model's reading of the conversation, not a legal finding. "Review" means a person should listen. No accuracy figure is published, because the labelled set behind the model is small and written by us. The white paper says what was measured and what was not.
Start
Put it on one line this week
Reception and finance first, the lines vishing targets. Route them through the hop, place one test call with the scam sample, read the verdict. Prepaid, per analyzed minute plus from 1.5 cents a call, $5 of credit at signup, no card. Already evaluating a vendor? Read how it compares with VeriSwitch SmartShield and FraudCentral VoiceShield.
Questions
Frequently asked questions
What is vishing?
Voice phishing: a phone call in which the caller pretends to be someone the person trusts, the bank's fraud desk, the IT help desk, a government agency, a supplier or the boss, and talks them into giving up a password, a one-time code, card details or a payment. It works because a call feels personal and urgent, and because the phone line has none of the filters e-mail has.
What is an anti-vishing solution?
Something that stops voice phishing rather than describing it. Training and "hang up and call back" depend on the person under pressure. Caller-ID apps and STIR/SHAKEN judge the number before the call. An anti-vishing solution for a business or a carrier reads the call itself: what the caller claimed to be, what they asked for, how they pushed; alerts a person with the reason; blocks the caller's next attempt; and keeps the recording and transcript as evidence. Open Voice Shield does this as one hop on the phone line.
Can vishing be detected automatically?
Yes, from the conversation. Open Voice Shield records each call on separate caller and callee channels, transcribes it, and a language model scores the transcript for the signatures of voice phishing: impersonation of an institution, a request for codes, cards, credentials or payment, threats and deadlines, remote-access instructions. In standard mode the verdict lands minutes after hangup with the red flags quoted from what was said, and the next call from a flagged number is blocked before it rings. Realtime mode, set up on request at a higher cost, scores the call while it is still up.
Does STIR/SHAKEN stop vishing?
No. STIR/SHAKEN authenticates the caller ID at signalling time, so a forged number can be flagged. It never hears the call. A scammer on a legitimately obtained number, or an "IT help desk" calling from a real mobile, passes it. The Canadian Centre for Cyber Security's guidance on vishing describes both the protection and its limit. Conversation analysis is complementary: it catches what the caller said.
Who is anti-vishing protection for?
Any organization that answers the phone: small and medium businesses whose reception, finance and support staff take calls; carriers and operators who carry the traffic and field the complaints; mobile operators who can sell scam call protection to subscribers as an add-on. The same hop serves all three. Only the routing differs.
What does it cost?
Prepaid, per analyzed minute at the plan rate plus one fee per call that covers transcription and analysis, from 1.5 cents a call. Signup carries $5 of credit and there is nothing to install: the phone system or carrier sends calls through the hop and the platform forwards them on.