展示 HN：看 AI 购物代理如何挑选推荐产品
英文标题：🔶 Show HN: See an AI shopping agent choose which products to recommend
这篇展示文章介绍了一款名为“AI 店面前台”的工具，当消费者使用 AI 代理购物时，代理依据结构化产品数据清单进行筛选，数据不完整的产品会被直接排除，而非降权。作者演示了上传产品目录后，因数据缺失而导致全数商品被拒之门外的场景，并估算由此造成的潜在年销售额损失。文章呼吁商家修复产品数据，并开放候补名单以获取自动修复建议，对依赖 AI 代理引流的电商开发者具有重要意义。
51.5 KB
中文
谁在 AI 店面的宾客名单上
在门口 · AI 店面有一份名单
世上最热门的店面
有个门童。
你的产品不在名单上。
当购物者派 AI 来买东西时，代理是拿着写字板工作的。这份名单就是结构化的产品数据。榜上有名的径直进入。缺少细节的名字会被拦下，你甚至还没意识到购物者来过。数据不完整的产品不是排名靠后，而是完全不被纳入考虑。 上传你的目录，逐个名字看看那些与你无关的销售是怎么发生的。
检查我的产品名单 观看示范门口
100% 浏览器内运行 · 你的目录绝不离开此页面 · 可从 SHOPIFY / WOO / MAGENTO 管理后台导出为 CSV
云雾消散 目录晴雨表
多数目录的读数在此。运行你自己的，看看所处位置。
门开着 · 今晚写字板上的名字
门关着。
一个都没进。
以下每个产品都在隔离绳处被拒之门外。合适的产品，合适的价格，只是不在名单上。没有争辩，没有申诉。今晚的花销跟着别人进去了。
▋▋▋ 每个产品被拒的原因已隐藏。加入下方候补名单，所有详情立即解锁。
估计流向隔壁的年消费额
$0
假设每个数据缺失的 SKU 每月产生一次代理引荐销售，并以你的标价计算，这是一个保守的下限，因为如今 AI 引荐的购物者转化率大约是其他渠道的两倍。
加入目录修复候补名单。
留下你的邮箱，我们会把你加入目录修复候补名单。准备就绪时，你损坏的产品会被自动修复，你将收到一份可重新导入的干净文件。
无法发送，请检查网络连接后重试。
分享我的结果 下载完整报告（CSV） 试用另一个目录
English / 原文
The Guest List who gets into the AI storefront
at the door · the ai storefront has a list
The hottest storefront in the world
has a doorman.
Your products aren’t on the list.
When a shopper sends an AI to buy, the agent works from a clipboard. The list is structured product data. Names on it walk straight in. Names with missing details get held back before you even know the shopper was there. Products with incomplete data aren't ranked lower. They aren't considered at all. Upload your catalog and watch, name by name, the sale that happens without you.
Check the list for my products Watch a sample door
100% IN-BROWSER · YOUR CATALOG NEVER LEAVES THIS PAGE · EXPORT AS CSV FROM SHOPIFY / WOO / MAGENTO ADMIN
FOG CLEAR CATALOG BAROMETER
most catalogs read here. run yours to see where you land
doors open · names on tonight’s clipboard
Doors closed.
none got in.
Every product below was turned away at the rope. Right products, right price, just not on the list. No argument, no appeal. The night’s spending walked in with someone else.
▋▋▋ The reasons each product was turned away are covered. Join the waitlist below and every line opens up instantly.
estimated annual spend walking in next door
$0
Assumes one agent-referred sale per dark SKU per month at your listed prices, a conservative floor, given AI-referred shoppers now convert at roughly twice the rate of other channels.
Join the waitlist for catalog repair.
Leave your email and we'll add you to the catalog repair waitlist. When it's ready, your broken products get fixed automatically and you get back a clean file to re-import.
couldn't send, check your connection and try again.
Share my result Download the full report (CSV) Try with another catalog