Amazon's new Interests feature turns your hobbies into AI shopping feeds

If you are deep into a niche hobby—whether it is tracking down vintage camera gear, hunting for specialized coffee tools, or keeping tabs on pickleball paddles—shopping online can feel like unpaid part-time labor. You are constantly refreshing feeds, tweaking search filters, and hoping the right item actually pops up.

Amazon wants to put an end to that endless searching. The company just introduced Interests, an AI-powered shopping feature designed to automatically track new products, restocks, and deals based on your hobbies.

Turning plain English into 24/7 store monitors

Instead of making you guess the right combination of keywords and category filters, Interests lets you set up persistent shopping prompts using conversational language.

You can keep it simple, like asking for "innovative kitchen tools and cookbooks to level up my cooking." Or you can get remarkably specific, typing out a prompt like: "I'm looking for wall art to decorate my home. I want something abstract or modern made of black metal, not canvas... under $100."

Behind the scenes, large language models translate your natural descriptions into structured queries and product attributes that Amazon’s existing catalog engine can parse. For instance, if you type "travel-friendly skincare products from premium brands," the model maps "premium brands" to specific label names in that department.

Once created, the prompt runs continuously in the background. Whenever matching inventory arrives, restocks occur, or relevant discounts drop, Amazon alerts you through a widget in the app's "Me" tab and directly on the home feed.

Availability and rollout

Right now, Interests is only accessible to a small subset of U.S. shoppers across the iOS and Android Amazon app, as well as the mobile web. Amazon plans to expand access to all U.S. customers over the coming months.

Those who have access can find it by heading to the "Me" tab in the app and looking for the Interests button. Shoppers can create multiple distinct prompts, and editing them is as simple as tapping a pencil icon to tweak budgets, materials, or stylistic preferences.

This marks another step in Amazon's generative AI push across retail, following its Rufus shopping assistant, AI Shopping Guides, and automated review highlights.

Why it matters

For consumers, Interests flips e-commerce on its head. Instead of forcing you to hunt down items every few days, it turns your long-tail hobbies into automated alerts. If you have distinct taste or a strict budget, letting an LLM watch the catalog for you makes immediate sense once it rolls out widely.

For the retail market, this shifts how discoverability works. E-commerce has spent decades refining keyword bidding and search optimization. When an LLM acts as the gatekeeper translating concepts like "aesthetic" or "premium" into search queries, merchant discoverability hinges on how well the model understands their brand identity.

We think Amazon's architectural choice here is quite clever. Rather than replacing the search interface with an intrusive chat assistant, it uses generative AI quietly as a semantic translator atop proven retrieval infrastructure.

The real risk lies in algorithmic bias. When an LLM translates qualitative prompts like "trusted brands" into hard database filters, smaller sellers and emerging labels risk being invisible if the model's training data does not already recognize them.