Leo AI Analytics
← All case studies
Case Study 01Food / Catering·Bay Area

The numbers were already there. Harnessing them revealed a 16% operating margin in 2 weeks.

A solo smash-burger operator running pop-ups, catering, and street fairs was already doing the hard part — tracking every event, every receipt, every shift. What was missing was a way to see what all that work was telling them. In 14 days, their own data became a real-time operations dashboard, an AI lead funnel, and proof the business was healthier than their ledger let on.

Operating profit revealed
$4,875
hidden by tax-ledger framing
Tax-prep time
3 days → 1 hr
reduced for first tax season after
Lane margin spread
13% → 64%
now visible across event types

The situation

Business X is a solo smash-burger operator doing weekly school lunch pop-ups, weekend street fairs, and occasional catering. Open ~14 months when we connected.

Revenue was real ($30k+ across 68 events), and the discipline was too — every event, purchase, and shift logged by hand across multi-tab Google Sheets. But the data wasn't organized to answer an operator's questions: inventory math happened in their head before every Costco run, catering leads sat in Gmail for 1-2 days, and a ledger built for tax season read like a $7k net loss. They weren't sure what was possible with the information they already had.

What we built

  • Operations dashboardPulled from their 2 years of spreadsheets. Revenue / expense / labor / cover KPIs, each clickable for full drill-down. Lane breakdowns showing their highest-margin lane ran at 64%.
  • Operating vs. tax-ledger P&LShowed them their ledger's -$7k loss included ~$11,900 in non-operating items (pass-through sales tax, equipment capex, business meals). Real operating P&L: +$4,875 (16% margin).
  • Opportunity CalculatorLane-aware. Type of event + headcount + price returns a take-it / skip-it verdict with cost breakdown and inventory shopping list — calibrated to their menu mix and the 17% prep buffer they themselves use.
  • Inventory tracker19 stock items with vendor info from their real purchase history. Real-time deficit alerts before each event.
  • Lead-funnel chatbot prototypeDemo of how AI would handle inbound catering inquiries — qualify the lead, check the calendar, auto-quote per their pricing rules, route to inbox with margin pre-calculated.

The insight that changed everything

The biggest moment in the project wasn't software — it was showing them what their own diligence had already earned. Strip the non-operating items out of the tax-season ledger and the food business runs a healthy 16% operating margin with real room to grow. The hard work was never the problem. It just needed to be concentrated where it moves the business most — and their numbers now show exactly where that is.

Pricing model

Free working prototype during discovery (~30 hours of build time on me). After they saw value and signed on, $75/month retainer covers hosting, monthly check-ins, bug fixes, and small features. Bigger items get quoted as flat-fee projects starting at $1,500.

What's next for them

Foundation is live. Greenlit the AI chatbot + Google Calendar integration as their first paid project. Evaluating an event-opportunity scraping agent next — AI that finds vendor openings in their radius and pre-fills the applications.

Want this for your business?

The same playbook applies whether you're a food business, a fab shop, a service business, or something I haven't built for yet. Book a 30-min call.