She built a Claude shopping assistant to stop buying cheap junk

Nicole Ruiz, a Brooklyn-based mother of two, describes the cognitive burden modern parents face navigating online purchasing — from filtering out Amazon knockoffs and drop-shipping brands to remembering artisan vendors discovered at farmers markets. To solve this, she created a dedicated Claude project (separate from her general AI queries) that stores a curated list of trusted heritage brands, along with explicit purchasing criteria: natural materials, multi-decade brand history, repairability, clear return policies, and avoidance of trendy direct-to-consumer brands with heavy influencer marketing spend.

The project instructs Claude to surface product recommendations in a standardized format — product name, photo, price, materials, care instructions, purchase link, and a note on the brand's trustworthy history. This last element proved especially valuable when Claude surfaced that one apparently appealing brand had been acquired two years prior and had since received consistently poor reviews. The system also flags AI-generated reviews and likely drop-shipping operations as part of its filtering logic.

Nicole demonstrates the flow live, searching for a can opener. Claude searches through vetted stores like Boston General Store — which itself pre-vets vendors — and surfaces the Noent Superkim can opener, a nearly 100-year-old brand, available on two websites ready for immediate purchase. She also shows how she uses a gift card balance as a shopping constraint, and how Claude surfaces heritage items like an LL Bean tote manufactured in Maine by the same craftspeople for over 80 years.

For returns, Nicole uses Claude Computer Use connected to her Gmail. She photographs a failed item (J.Crew pants that wore through in four months), describes the problem verbally using Whisper Flow, and asks Claude to find the original receipt, extract the order number and item details, and draft a refund request email. The drafted email includes all relevant details and frames the request around the brand's quality promise rather than just their return window — a tactic she says often works because manufacturers sometimes already know about production defects and are waiting for customer complaints.

Nicole argues this system ultimately benefits small artisan brands and heritage manufacturers whose outdated websites put them at a disadvantage against Amazon, by allowing AI to surface them based on quality signals rather than SEO or ad spend. She frames the entire project not as replacing human parenting judgment but as automating the administrative digital labor so more time can be spent on direct human interaction with family.