How I AIShe built a Claude shopping assistant to stop buying cheap junk
CHAPTERS
Nicole’s goal: automate shopping admin to buy fewer, better things
Nicole Ruiz explains how parenting today involves constant online “administrative tasks” (buying, returns, support emails) that create mental overload. She uses Claude to offload that work so she can focus on the human parts of parenting while improving the quality of what enters their home.
Why Amazon-style convenience leads to cheap junk and bad decisions
Nicole describes the common cycle: panic-buy an item quickly (often on Amazon), then regret the quality, materials, or authenticity. She wants thoughtfully made, repairable products and less exposure to ads, SEO spam, and drop-shipped knockoffs.
From Apple Notes to a Claude Project: building a household purchasing system
Nicole consolidates her trusted vendor list (previously in Apple Notes) into a dedicated Claude Project with persistent instructions and memory. Keeping it in a separate project prevents other Claude conversations from “overfitting” to shopping constraints and keeps her household system organized.
Purchase criteria that Claude must enforce (quality, returns, anti-drop-shipping)
The project instructions encode Nicole’s values: heritage brands, vetted retailers, returnability, avoidance of trendy DTC brands, and skepticism of AI-generated reviews. Claude is asked to look for signals like durable materials, repair options, and whether a business truly stands behind its products.
Output formatting: make the invisible checklist visible
Nicole standardizes how Claude returns recommendations so she can decide quickly. Results should include product name, photo, price, materials, care instructions, purchase link, and a short explanation of brand trustworthiness (including changes like acquisitions that degrade quality).
The “anti-to-do list”: using AI to reduce repeated mental overhead
Claire highlights the broader pattern: many household tasks involve an internal checklist (fiber, delivery timing, returns, resale) that repeatedly consumes attention. Capturing that checklist once inside a Claude project reduces ongoing cognitive load across frequent purchases.
Live demo: shopping for a can opener with Claude’s web search
Nicole runs the system on a simple, failure-prone household item: a can opener. Claude searches preferred vendors first and returns comparable options with heritage context, pricing, and quick links—making it easier to choose quality without spending time browsing.
Feedback loop: telling Claude what you bought (and why)
Nicole sometimes reports back what she purchased so the project improves over time and surfaces better vendors. Clothing remains harder due to sizing variability, but Claude can still help interpret inconsistent brand size guides.
AI as an equalizer for heritage brands with terrible websites
Nicole notes that many century-old, high-quality manufacturers have outdated web UX, which makes purchasing from them slower than using Amazon. Claude reduces that friction, helping consumers discover and buy from these brands without fighting poor site design.
Returns and refunds with Claude Cowork: turning photos + email into action
When items fail (e.g., kids’ pants wearing through quickly), Nicole uses Claude Cowork connected to Gmail to find receipts, item numbers, and order details, then drafts a strong refund email. Voice input and mobile capture make it easy to act immediately rather than letting returns pile up.
Strategic shopping: using gift cards and budgets to pick “classic” items
Nicole uses Claude to optimize constrained spending—like a gift card balance—while still adhering to her quality rubric. Claude often surfaces a brand’s most enduring, best-known products and provides craftsmanship and provenance details that reinforce trust.
Vetting unfamiliar brands: legitimacy checks, scaling signals, and review forensics
Nicole asks Claude to analyze brands she encounters via ads to determine if they match her standards or are likely low-quality. The system considers signals like heavy influencer marketing, private equity/rapid scaling issues, inconsistent manufacturing, leadership controversy, and review patterns.
Recap: the end-to-end ‘buy better’ loop (recommend → purchase → enforce quality)
Claire summarizes Nicole’s workflow: maintain a vetted vendor list and criteria in a Claude Project, use it for recommendations and brand checks, then use Claude Cowork to execute returns/refunds when products don’t deliver. The result is fewer low-quality items in small spaces and less ongoing admin work.
Lightning round: parenting impact and how to correct Claude when it’s wrong
Nicole explains the parenting benefit: reclaim time and attention by automating “email-job” chores. When Claude misses the mark, she recommends direct feedback—ask what went wrong, then revise criteria like you would when managing an employee.
