
Andrew Bell joined Scott Needham on the Smartest Amazon Seller podcast for an episode called "How to Rank #1 in Amazon's AI Shopping Results," a research-and-playbook deep dive built around ReFiBuy's 100,000-prompt Alexa for Shopping study. It's exactly the kind of practical, data-first conversation we like here, so we wanted to bring it to you.
Andrew walks through how Alexa for Shopping actually finds products: it fans out into multiple Amazon searches, pulls in a pool of candidates, and re-ranks them on context and personalization. From there, he and Scott dig into what the study found, including why a 4.0-star rating acts as a floor, how semantic shelves change what "position one" really means, and why use cases, titles, and item highlights carry so much weight. They also cover how to optimize for AI visibility without giving up search demand, how the biggest brands on Amazon are approaching it, whether ROI can be measured yet, and whether off-Amazon content is starting to shape the answers.
Watch the full video on YouTube below:
Highlights
𝗦𝗲𝗮𝗿𝗰𝗵 𝘀𝘁𝗶𝗹𝗹 𝗳𝗲𝗲𝗱𝘀 𝘁𝗵𝗲 𝗮𝗴𝗲𝗻𝘁: Alexa for Shopping fans out into multiple Amazon searches, pulls in a pool of 40–50 products, then re-ranks them on context and personalization. Amazon SEO hasn’t gone away; you need both.
𝗦𝗲𝗺𝗮𝗻𝘁𝗶𝗰 𝘀𝗵𝗲𝗹𝘃𝗲𝘀: The bold headings in an answer each represent their own set of queries. Being #1 on your shelf can matter more than #1 overall. “You don’t want to be position one for an irrelevant attribute.”
𝗨𝘀𝗲 𝗰𝗮𝘀𝗲𝘀 𝘄𝗶𝗻: 53% of product rationales pointed to a real-world use case. Item name, item highlights, and the first bullet carry the re-rank.
𝗗𝗼𝗻’𝘁 𝘀𝗮𝗰𝗿𝗶𝗳𝗶𝗰𝗲 𝘀𝗲𝗮𝗿𝗰𝗵: “As soon as you start sacrificing core search demand, you’ve already lost your qualified visibility.”
𝗠𝗶𝘀𝘀𝗶𝗼𝗻 𝗺𝗮𝗽𝗽𝗶𝗻𝗴: Break your Search Query Performance terms into customer missions, then check which semantic shelves match each one.
𝗧𝗵𝗲 𝗿𝗲𝗽𝗼𝗿𝘁 𝗲𝘃𝗲𝗿𝘆𝗼𝗻𝗲 𝘄𝗮𝗻𝘁𝘀: No one has Alexa-attributed sales yet. Sponsored ads prompt reports are the closest proxy, and they show plenty of impressions but few attributed sales.
Thanks to Scott for having Andrew on and for asking the questions sellers actually care about. Enjoy the conversation.
🎧 Listen wherever you get your podcasts: Spotify, Apple, etc.
We’re here to make sense of brand strategies for agentic commerce on Amazon, Walmart and beyond. Subscribe, share, and stay tuned!



Fanning out into several searches and then re-ranking explains why classic keyword work still matters underneath. Search first, then context, is a useful order to remember.
The fan-out into 40-50 search hits before the re-rank is the part I'd sit with. If use cases and the first bullet drive that re-rank, a quick check is to ask Alexa the same mission three ways and see whether your opening line still matches what gets pulled. Search demand stays the pool; the agent just reorders it.