Claude to build a kitchen appliance buying agent
Yoni Goyhman used the last week of paternity leave to build a shopping assistant for his kitchen renovation. Claude Code produced a system with two connected flows: translating broad appliance needs into product criteria, then comparing prices for a selected shopping list.
linkedin.com
Suggested prompt
Suggested prompt written by Agents In Use, not the source author. Replace bracketed inputs and review tool permissions first.
Build a kitchen appliance buying assistant with separate product discovery and seller comparison. Let me select the basket, find sellers carrying several items, and draft a combined discount request for me to send.
Reach people exploring AI agents
See placements →What happened
Goyhman reports purchasing kitchen appliances with a 7% bulk discount and ₪8,000 saving; independent purchase receipts are not supplied.
Reported by Yoni GoyhmanHow the agent did it
- Turn household appliance needs into explicit product criteria and a selected list.
- Find sellers with overlapping products and draft a combined discount request.
- Inspect the results, send the request and review the application's changes before deployment.
Where a person stepped in
Goyhman chose products, used the buying system and increasingly reviewed every commit while stopping automatic production deployment.
Yoni Goyhman used the last week of paternity leave to build a shopping assistant for his kitchen renovation. Claude Code produced a system with two connected flows: translating broad appliance needs into product criteria, then comparing prices for a selected shopping list.
The discovery flow searched for matching models and let the user choose items. The price flow looked for sellers carrying several of those products, grouping the purchase to support a bulk discount request. It collected seller WhatsApp numbers and drafted messages that users could send directly. A back-office view retained seller data and exposed the agent's activity for inspection.
Goyhman says he used the application to buy his kitchen appliances, obtained a seven-percent bulk discount and saved ₪8,000. Those are his reported results; the article does not supply receipts establishing a market-wide cheapest price.
The build also became harder to maintain. Bugs and regressions accumulated, including unreadable interface text. Goyhman began reviewing every commit manually and explicitly told Claude to stop automatic production deployment. The useful shopping pattern is selecting a basket before searching for overlapping sellers and preparing a joint discount request. The account pairs a real buying outcome with the owner's increasingly strict supervision of the software that enabled it.
Prompt
Suggested prompt (written by Agents In Use, not the author's)
Source: www.linkedin.com · Yoni Goyhman
Frequently asked questions
How did the app support a bulk discount?
It identified sellers carrying multiple selected items and drafted a combined request using collected WhatsApp contacts.
Did it send negotiation messages without the user?
The described flow drafts requests and lets users send them directly.
What changed in supervision as the app grew?
Bugs and regressions led the author to review every commit manually and tell Claude to stop automatically deploying production changes.
More details
About this account
Claude built the product-discovery, seller-comparison and negotiation-drafting application.
Firsthand source opened and read. Suggested prompt written by us from the reported workflow; no original task prompt is quoted.
Credit: Yoni Goyhman
Source published 5 Feb 2026
Source checked 14 Oct 2026
Tools and topics
Claude Code · Shopping agent · Seller websites · WhatsApp
Kitchen appliances · Bulk discounts · Seller comparison
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.