How Amazon lets customers check if a message is really from Amazon

Amazon impersonation messages work because almost no one can tell a fake delivery alert or account warning from a real one. According to the Global Anti-Scam Alliance, Amazon now lets customers check a suspicious message against the company's own records rather than guess based on how it looks.

The scale of the problem explains why. In the United States, impersonation scams were the most frequently reported fraud category in 2025, making up nearly one in three fraud reports and $3.5 billion in reported losses, according to the Federal Trade Commission. Amazon says 360,000 customers contact its customer service each year just to ask whether a message claiming to be from Amazon is real.

What the scam depends on

An impersonation message only works if the person reading it has no way to check it against anything. A text claiming to be from Amazon looks plausible because it borrows Amazon's name, logo and tone. The reader is left to judge authenticity from the message itself, which is exactly the thing a scammer controls. That is the gap Amazon's system is built to close.

Instead of scanning a message for suspicious links or odd wording, the system compares what the customer submits against Amazon's own records of outgoing communications, records the company says cover billions of messages. If the message matches something Amazon actually sent, it is confirmed. If it does not, the customer is told so. If a match cannot be established either way, the customer gets guidance on what to do next, along with relevant security advice.

This is a different method from most scam detection, which looks for signs that something is wrong with the message. Amazon's approach instead asks whether the claimed sender actually sent it. That shifts the question a customer has to answer from "does this look fake?" to "did Amazon send this?", which is a question Amazon, not the customer, is positioned to answer.

Three ways in, one shared outcome

Amazon offers three channels for this verification, so a customer can use whichever suits the situation. Every submission made through Alexa for Shopping is also automatically passed to Amazon's customer protection and enforcement team. That means each individual check also becomes a data point in Amazon's wider effort to track impersonation attempts, feeding enforcement rather than ending once the customer gets their answer.

The mechanism the scam relies on is a one-way message with no way to check it. Verification against the real sender's own records removes that one-way quality: the customer is no longer asked to spot the fake, only to confirm whether the real company sent it.