Amazon’s shopping AI can now verify whether suspicious messages really came from Amazon

According to TechCrunch, Amazon said that about 360,000 customers each year contact its customer service department to ask whether a message they received from Amazon is real or a scam. To answer such questions, Alexa for Shopping will compare the message against a record of every message Amazon has sent globally, while also analyzing sender information, content, timing, formatting, and metadata. Amazon said the AI can “definitively” confirm whether a communication originated from its own systems, and the system will improve as customers report more suspicious messages.

The Verge reported that Alexa for Shopping will only confirm a message is genuine if it is “completely certain.” For example, a customer might ask whether Amazon sent a one-time password by text at a certain time, and the assistant can confirm that it was authentic. If the AI determines the message is not from Amazon, it will tell the customer to check their orders in the app and contact Amazon’s support team directly.

This is not Amazon’s first attempt to address the problem. Earlier this year, the company launched the email address verify@amazon.com, where anyone can forward suspicious messages and receive confirmation of whether they are real, and it also offers a similar online form on its customer service site. The new AI capability builds on those efforts, adding a more immediate way for customers to get answers.

The update is part of Amazon’s broader shift to focus Alexa for Shopping on practical, shopping-related help. The assistant already can find personalized deals, create shopping guides, help reorder everyday items, offer AI product overviews, and transcribe handwritten shopping lists. On Tuesday, Amazon added an “Update Me When” feature that lets customers set alerts for new product releases from favorite brands, authors, or artists.

Alexa for Shopping is available through both Amazon’s website and its mobile app.