The short answer
On any call that comes to you, never give out a password, a sign-in or verification code, or customer information — no matter who the caller says they are. Take their name and company, end the call politely, and call back on a number your business already has on file. If the request was genuine, it survives a callback. If it was not, the callback is what saves you.
Why this works. A stranger's whole advantage is staying on their call, on their timeline. Hanging up and calling a number you trust takes both away. Real banks, real IT support and real suppliers can all be reached on numbers you already have — and none of them need your staff to read out a code.
The three things that are never given out on an inbound call
Post these three at every phone. They are short on purpose — a rule staff have to remember under pressure has to fit in one breath.
- Passwords. Not the Wi-Fi password, not the email password, not “just the shared one everyone knows”. A password read out once cannot be unread, and the caller can use it from anywhere.
- Sign-in and verification codes. The six-digit code that arrives by text or app is the login, for whoever holds it. Reading it out is the same as handing over the key while someone waits at the door.
- Customer information. Names, phone numbers, appointment times, who owes what, whether someone is a customer at all. Each small fact makes the next call — to you or to that customer — sound more convincing.
Caller ID is not proof. The name and number on the display can be faked (spoofing). A call can show your bank's real name and still come from a stranger. Judge the request, not the display — and never judge by how confident or friendly the voice sounds. Confidence is the tool of the trade.
The callback rule, in five steps
One procedure for every unexpected request, whoever it claims to be from. Staff do not have to decide whether a caller is lying — the procedure decides.
- Pause. Nothing about a password, a code or a customer's details is ever urgent on someone else's call. Urgency is a pressure tool, not information.
- Take the details, give nothing. Write down the caller's name, the company they claim, and what they say they need — and give out none of the three things above, not even to “confirm” them.
- End the call politely. Use the script below. You are not accusing anyone; you are following the house rule, and the house rule applies to every caller equally.
- Call back on your number. Use the number on the business's own card, statement, contract or the official website you type in yourself — never a number the caller gave you, and never by redialling the incoming call.
- Log it. Date, time, claimed name and company, what was asked, and what you did. Two lines in a notebook by the phone. Patterns only become visible when calls are written down.
A word-for-word script for the front desk
These are templates — adjust the wording to your business, but keep the shape: thank, state the rule, take the details, promise the callback, end the call. The rule is always “our policy”, never a personal refusal, so no caller can argue one employee into an exception.
“From the bank” — asking for a code or password
“Thanks for calling. Our policy is that we never give out passwords or codes on a call that comes to us. I'll take your name and what this is about, and we'll call the bank back on the number on our own statement. What's your name and department?”
“From IT support” — asking you to sign in or read a code aloud
“I understand. Our policy is that sign-ins and codes only happen on a call we place ourselves. Let me note your name and ticket number, and we'll call our support number back. If you set this up with us, it'll be in our records.”
“From a supplier” — asking about a customer, an order or an account detail
“I'm not able to share customer or account details on an incoming call — that's a house rule for everyone. If you leave your name and company, the right person will call you back on the number we have on file.”
If the caller pushes back — gets annoyed, threatens a fee, says the account will be locked, or asks for “someone who can actually help” — that is the strongest sign the call is wrong. Repeat the rule once, word for word, and end the call. A genuine caller can be re-reached in two minutes; a stranger needs you to stay on this call.
The five pressure lines, and what they really are
- “This is urgent — the account will be locked today.” Manufactured deadline. Real lockouts are handled on the callback, in minutes.
- “I'm calling from head office / the owner asked me to.” Borrowed authority. A claim is not proof; check with the owner on a channel you already use.
- “You helped me with this last week.” Borrowed familiarity. Small true details (your name, the owner's name, a supplier's name) are easy to find and prove nothing.
- “I just need one small thing.” The small ask first. One “harmless” detail today is the convincing detail in tomorrow's call.
- “Don't tell your manager, it'll reflect badly.” Isolation. Any caller who asks for secrecy from your own team has answered the question for you.
If something was already given away: the first hour
It happens to careful people. What matters is the next hour, not the blame. General information only — your own accounts and policies come first.
- Tell the owner now. Immediately, before anything else. A password can be changed in minutes; a secret kept for a day cannot.
- Change the password from a clean device — a device you trust, signed in the normal way. Never “fix” it by following a link or number from the suspicious call.
- Sign out other sessions where the account offers it, and turn on two-factor sign-in if it was off.
- Check the recovery settings — the recovery email and phone on the account should be yours, and only yours.
- Treat any code that was read out as used, even if nothing has happened yet, and watch the account for sign-ins or changes you do not recognise.
- Write the timeline down — what was asked, what was shared, when, and what you changed. If you later need your bank's or provider's help, this page is what they will ask for.
One-page checklist for the front desk
- Never give out passwords on an inbound call — to anyone, for any reason.
- Never read out a sign-in or verification code. The code is the key.
- Never share customer names, contact details, appointments or account facts.
- Caller ID is not proof — names and numbers can be faked.
- Take the caller's name, company and request. Give nothing back.
- End politely with the house rule: “We call back on our own number.”
- Call back only on a number from your own records — never one the caller gave.
- Log every such call: date, time, who claimed to call, what was asked.
- Something shared already? Tell the owner at once — speed beats blame.
Rules on paper are one thing. What your team actually says is another.
Most owners have never heard what their front desk says to a confident stranger — the risk stays invisible until a real caller uses it. HumanCheck tests exactly that, with your signed written permission first: a real caller phones and messages your staff the way a stranger would, every scenario stops at the moment of disclosure or refusal, and anything disclosed is never used, entered or kept. You get a written report by role — never by name — with fixes like the ones on this page, plus a 15-minute walkthrough call.
Standard Check $149 one-time (up to 10 staff, 3 call + 2 message scenarios) · Deep Check $299 one-time (up to 25 staff, 5 call + 4 message scenarios, follow-up, briefing sheet, one re-check within 60 days). There is no online checkout, by design — the signed authorization comes first.
See how HumanCheck worksCommon questions
What should an employee do when a caller says they are from the bank and asks for a sign-in code?
End the call politely and call the bank back on the number printed on the business's own card, statement or the bank's official website — never a number the caller provides. Banks and real IT support do not need your staff to read out a sign-in code; the code is the key itself. Use the script on this page: take the caller's name and company, say you will call back, and hang up first.
Why doesn't caller ID prove the caller is genuine?
The name and number shown on the phone can be faked — this is called spoofing, and it is cheap and common. A call that displays your bank's real name and number can still come from a stranger. That is why the rule is based on calling back a number you already trust, not on how the incoming call looks or sounds.
What are the three things staff should never give out on an inbound call?
Passwords, sign-in or verification codes, and customer information (names, contact details, appointment or account details). All three can be handed over in seconds by a helpful employee, and all three are exactly what a stranger needs to get in or to build the next, more convincing call.
A caller says the owner asked them to call. Should staff help them?
Not on that call. 'The owner asked me to' is a claim, not proof — and it is one of the most common pressure lines. The safe move is the same callback rule, plus a quick check with the owner through a channel you already use. A genuine caller working for the owner will have no problem with being called back.
An employee already read out a password or a code. What happens now?
Tell the owner immediately — speed matters more than blame. Change that password from a clean, trusted device, sign out other sessions where the account offers it, check the account's recovery email and phone, turn on two-factor sign-in, and write down a short timeline of what was shared and when. Treat any code that was read out as used, even if nothing seems to have happened yet.
Is this guide legal or security advice for our specific business?
No. These are general front-desk habits for a small team, written in plain language. Regulated businesses — health, finance, legal — may have additional rules about calls, recordings and customer information; check your industry's requirements or ask a professional in your field.