Digiday: Amazon’s shopping agent now takes orders by text, with Sponsored Product ads in the replies

Digiday: Amazon’s shopping agent now takes orders by text, with Sponsored Product ads in the replies

Digiday reports that Amazon opened Alexa for Shopping by text to a small, invite-only group of U.S. customers on November 6, 2026. The agent can search and compare products, suggest options in a text thread, and place an order when the customer chooses. Sponsored Products may appear among those suggestions; purchases the agent completes on its own remain ad-free. Amazon is not charging advertisers for these placements yet. The placement is limited to the discovery stage, while a customer is still choosing. Digiday presents that boundary as important: putting an ad into an agent-made purchase could mean the customer receives a product they would not have chosen. As an Escape Meridian interpretation of the article’s customer-choice concern, keeping ads in suggestions rather than agent-made purchases preserves a distinction between influencing a choice and making it. The text channel also gives Amazon a way to test how much advertising people will accept in personal conversations, though the article does not report customer responses to the test. Independent ad tech consultant Jonathan D’Sousa-Rauto says the definition of an impression and the reporting and measurement will matter in judging whether the ads improve efficiency for brands. That leaves an open question: how will Amazon define and measure an impression in these conversations, and what evidence will show whether the

Publication: a876ed7f-437e-4ad9-8a38-b702c70f3567

attribution — pass: The reporting is correctly attributed to Digiday (byline Krystal Scanlon). The impression and measurement point is correctly attributed to independent ad tech consultant Jonathan D’Sousa-Rauto. The discovery-only boundary is presented as Digiday's framing, and the extra inference is labeled as Escape Meridian interpretation.

said independent ad tech consultant Jonathan D’Sousa-Rauto

summary — uncertain: The summary breaks off mid-sentence at the end of its third paragraph (“what evidence will show whether the”). A reader sees an unfinished thought. Smallest repair: complete or remove that sentence, since the question field already carries it. Two central terms are also not explained in plain language: “impression” (the core of the question) and “Sponsored Products.” A reader outside ad tech may not follow why measurement matters. Suggest adding both to the term library. Otherwise, the summary faithfully conveys the actor, the development and the customer-choice stake.

the definition of an impression is critical as well as the reporting and measurement

date — pass: originalPublished (2026-11-06T14:00 UTC, publisher page) matches the article byline date of November 6, 2026. The draft's statement that the feature went live on that date matches the source's “went live today.”

By Krystal Scanlon • November 6, 2026

support — pass: The source states each of these points: the invite-only U.S. launch, ads in the suggestions, ad-free agent purchases, no charge to advertisers yet, a discovery-only boundary, and a test of ad tolerance. The Escape Meridian interpretation is labeled, has a supporting excerpt and stays conditional. The limitations correctly say the source reports no customer responses or efficiency results. The question follows from D’Sousa-Rauto's remark and is not a leading claim.

which gives it a cheap way to test how much advertising people will put up with in a text thread

media — pass: The selected media is the verified direct link to the original Digiday article. That is the most direct way for a reader to reach the source.