Replacement Parts Kits for Vintage Small Appliances

How the agent runs it

Each morning a sourcing agent queries the Airtable ledger for kit SKUs whose stock fell below par, checks real supplier prices against the stored cost-basis record, and generates draft purchase orders only if the computed margin on the restocked kit exceeds a hardcoded 38% floor stored in the ledger — it never relies on memory for prices or inventory counts. An order-fulfillment agent polls Shopify for new paid orders, confirms physical inventory count in ShipStation before marking anything as 'in stock', prints a packing checklist the warehouse operator follows, and writes a completed-order record back to Airtable with exact COGS pulled from the BOM table. A nightly review job computes actual margin per kit from the ledger, appends a dated lessons entry to a shared context file, and flags any order where realized margin deviated more than 5 points from plan for next-day human review.

Who this is for

Best suited for someone with a light hardware or e-commerce background — perhaps a former Etsy seller or a hobbyist restorer — who already understands appliance parts sourcing and can negotiate with two or three suppliers in the first month. The owner needs to be comfortable reading an Airtable dashboard daily and acting on escalation tickets within 24 hours; this is not a passive-income play. A person who enjoys the physical side (spot-checking kit contents, approving purchase orders) will find the 19% of manual work satisfying rather than burdensome.

Market opportunity

The repair-economy has accelerated since 2022 EU right-to-repair legislation and US executive orders encouraging appliance longevity; Google Trends for 'KitchenAid repair parts' and 'vintage mixer restoration' have grown 34% YoY through 2025. The addressable buyer is a DIY restorer who finds sourcing eight individual parts across three suppliers frustrating — a pre-packed kit with a clear BOM solves that friction. Margins are defensible because no dominant competitor has productized per-model kits at this price point; most sellers list parts individually on eBay with no curation.

Tech stack

Monetization

Price: $38 for a single-appliance basic kit (e.g. KitchenAid stand mixer carbon brushes + gaskets); $65 for a premium multi-part restoration kit. Volume discount capped at 10% for orders of 5+ kits, enforced at the Stripe price-table level — no agent override. Variable cost per customer: parts COGS $16.50 (average across kit tiers, from Airtable BOM) + poly mailer + tissue + label $1.80 + Stripe fee 2.9% + $0.30 ≈ $2.30 on a $48 average order + ShipStation postage $5.40 + Claude API per order ≈ $0.18 + human support time 4 min at $25/hr ≈ $1.67. Total variable cost: $27.85 on an average $48 order. Fixed monthly costs: Shopify Basic $39 + Airtable Team $20 + ShipStation Starter $25 + Claude API base + sourcing-agent runs ≈ $55 + parts pre-stock working capital interest (estimated) $30 + business insurance $40. Total fixed: $209/mo. Break-even: $209 fixed ÷ ($48 − $27.85) gross margin per order = $209 ÷ $20.15 ≈ 11 orders/mo to cover fixed costs. Margin at target: at 110 orders/mo (revenue ≈ $5,280), gross profit = 110 × $20.15 = $2,217 minus $209 fixed = $2,008 net operating income, a 38% net margin. This assumes the agent makes pricing errors on roughly 5% of orders (corrected by Stripe floor), costing an estimated $55/mo in margin give-back already baked into the variable cost figure.

Key risks

• → Over-discounting (Vend failure mode): the agent could stack a welcome-code with a bundle discount and sell below cost. Control: Stripe price-table rules enforce a hard minimum charge per kit SKU; no coupon code can reduce the final charge below that floor regardless of what the agent communicates to customers.
• → Fabricated inventory (Vend failure mode): agents may mark a kit 'in stock' before a supplier shipment is received. Control: ShipStation's physical scan is the only write path to the inventory ledger; the agent is blocked from marking items available until the ShipStation webhook fires a 'received' event and the Airtable record updates.
• → Adversarial customers claiming bulk-discount authority (Vend adversarial failure): a buyer emails claiming they are a 'reseller partner' entitled to 40% off. Control: no discount tier above 10% exists in the Stripe price table at all; the agent cannot grant it regardless of the claim, and any request invoking special authority escalates to a human-review ticket automatically.
• → Cost-basis drift over long horizons (Vend drift failure): supplier prices change and the agent keeps quoting kits using a stale cost figure, eroding margin invisibly. Control: the sourcing agent must read the supplier's live price API or scrape at the time of each purchase order and write the new cost-basis to Airtable before computing margin; it may not use any cost figure older than 48 hours.
• → Authority impersonation (Vend adversarial failure): an email arrives claiming to be from a supplier requesting a bank-account update for invoice payment. Control: supplier payment details are stored in Airtable and can only be changed by a human operator through a separate admin interface; the agent has read-only access to payment records and is structurally unable to update banking details.

Getting started

1. 1
   Build the Airtable BOM and cost-basis ledger first

   Before writing a line of agent code, manually enter 10 starter kit SKUs with supplier, part numbers, unit cost, and target sell price into Airtable. This ledger is the source of truth the agent will read — getting it right before automation touches it prevents fabricated-cost errors from day one.
2. 2
   Set Stripe price-table minimums before any agent goes live

   Create a Stripe price table with a hard minimum charge per kit SKU (e.g. $34 floor for basic, $58 for premium) and set coupon restrictions so no code reduces the charge below that floor. This one-time configuration replaces the need to trust a system prompt to prevent below-cost sales.
3. 3
   Wire ShipStation as the sole inventory write path

   Configure ShipStation so that inbound supplier shipments require a physical scan before the webhook updates Airtable stock counts. Test this with two dummy receipts to confirm the agent cannot read 'in stock' until the scan fires — this eliminates the fabricated-inventory failure mode.
4. 4
   Deploy the sourcing agent and run it shadowed for two weeks

   Let the sourcing agent generate draft purchase orders but route all of them to a human-review Slack channel for manual approval before any money moves. Review each draft to calibrate the 48-hour cost-basis staleness rule and confirm margin calculations match your Airtable figures before granting the agent autonomous PO authority up to $200.
5. 5
   Define and test all five human-escalation triggers before launch

   Hard-code triggers in the agent scaffolding: any refund request over $15, any supplier payment-detail change attempt, any order with realized margin below 25%, any inbound message containing the words 'reseller', 'wholesale', or 'partner discount', and any daily margin deviation flag from the nightly review job. Confirm each trigger fires a Slack alert and logs a ticket in Airtable before you process your first real customer order.

// done for you

Want us to build
Replacement Parts Kits for Vintage Small Appliances
for you?

We contract experienced engineers to deploy AI agent businesses end-to-end — custom domain, branding, live and earning in weeks. No code required on your part.

We reply within 1 business day · No obligation · Canadian-based team

Related ideas