Live work · ProfitPilot
ProfitPilot.
Embedded Shopify app. True profit per order with Claude P&L summaries.
Most Shopify analytics apps show revenue. This one shows profit, per order, with ad spend, COGS, and platform fees allocated correctly.
I · About
Shopify merchants run their stores on dashboards that lie to them. Revenue looks great until you back out the Meta ad spend, the COGS, the platform fees, the shipping subsidies. ProfitPilot runs the math right: every order gets its true profit calculated, then Claude writes a plain-English P&L summary the founder can read in 30 seconds. Built embedded in Shopify Admin so it lives where the work happens.
II · How it works
The pipeline.
- 01OAuth into Shopify, fetch orders + variants + customers
- 02OAuth into Meta and Google Ads, fetch ad spend at the campaign + creative level
- 03Per-order expense allocation: ad-attribution by UTM + platform conversion data
- 04Daily background job recomputes P&L with current cost of goods + shipping subsidies
- 05Embedded Polaris dashboard: today, this week, this month, all-time
- 06Claude API generates a daily P&L narrative: 'You did $X in revenue, $Y in profit. The biggest profit driver was Z. Here are three things to look at.'
- 07Stripe billing for Pro tier ($29/mo) and Growth tier ($59/mo)
III · Sample
Why this is the production build, not the toy.
Most AI side projects ship a chatbot wrapped around an API call. ProfitPilot ships an embedded Shopify app with OAuth, Stripe billing, per-customer data isolation, daily background jobs, and a domain model that requires understanding accounting. The Claude piece is one part. The other 80% is real software: schemas, migrations, billing webhooks, Shopify App Store submission, customer support flows. This proves the AI workflow ships to market, not just to GitHub.