You Don't Know What Each Dish Actually Costs You

The Real Problem

You run a busy Auckland restaurant. Revenue is good — covers are up, the dining room is full most nights. But at the end of the month, when you sit down with your accountant, the profit is... thin. Again.

NZ hospitality runs on razor-thin margins. Even a very well-run restaurant struggles to net 20% profit after all overheads. Labour costs have hit a record 40% of revenue in 2025. That leaves food cost as the one lever you can actually control — but most single-location NZ restaurants don't have a system to track it in real time.

Here's what typically happens: you set menu prices when you launch a dish, based on a rough estimate of ingredient costs. But ingredient prices change — sometimes weekly. Your salmon supplier raises prices by 15%. Butter goes up. The seasonal produce you were using runs out and you substitute something more expensive. Each change is small, but over a month, your food cost percentage creeps from 28% to 34% without you noticing.

You find out at the end of the month. By then, you've served 2,000 covers at the wrong margin.

Enterprise tools for food cost management exist — MarketMan, meez, WISK, Apicbase. But they're designed for multi-location chains, priced in USD, and way too complex for a single-location NZ cafe or restaurant. So most NZ owners do food costing on a spreadsheet. Or instinct. Or not at all.

Why Existing Tools Don't Solve This

Your POS (Lightspeed, Square) tracks what you sell, not what it costs to make. Xero tracks your total food purchases, but not per-dish costs. The enterprise recipe costing tools are overkill — you don't need a system designed for a 50-location chain when you run one restaurant.

What you need is something that sits between "no system" and "enterprise food management platform" — something that works the way you actually work.

How AI Solves This

When your supplier invoices arrive — by email, as a PDF, or on paper — you forward or photograph them and send to your AI assistant on WhatsApp:

[Photo of Gilmours invoice] "This week's dry goods order"

The AI reads the invoice (OCR + AI extraction), identifies the items, and updates your ingredient cost database:

Invoice processed: Gilmours — 12 Mar 2026

📈 Price changes detected:

• Calrose rice: $38.90/10kg → $42.50/10kg (+9.2%)
• Olive oil (extra virgin): $24.99/2L → $27.50/2L (+10.0%)
• Butter: $7.49/500g → $7.49/500g (no change)

Impact on your dishes:

• Risotto: food cost now $6.20 (was $5.75) — margin dropped from 72% to 69%
• Pasta aglio e olio: food cost now $3.85 (was $3.50) — margin dropped from 81% to 79%

No menu price changes recommended yet — changes are within tolerance. I'll flag if cumulative increases exceed 5% on any dish.

Over time, as you process more invoices, the AI builds a complete picture of your ingredient costs — and can tell you exactly what each dish costs to plate, in real time.

Monthly Margin Report

At the end of each month:

Food Cost Report — February 2026

Overall food cost: 29.3% (target: 28%) Highest margin dish: Mushroom arancini (82% margin, $2.40 food cost) Lowest margin dish: Scotch fillet (48% margin, $18.50 food cost)

Dishes exceeding cost threshold:

• Seafood platter: food cost up 12% since pricing — recommend price review
• Duck confit: duck price increased 8% — still within tolerance but trending up

Supplier comparison: Your average protein cost from Supplier A is 7% higher than Supplier B for equivalent items. Consider shifting lamb and chicken orders.

How It Works Under the Hood

1. You photograph or forward supplier invoices via WhatsApp
2. AI extracts line items, quantities, and prices using OCR
3. Costs are matched against your RECIPES.md workspace file (your dish recipes with ingredient quantities)
4. A food-cost skill calculates per-dish costs and margin percentages
5. Price changes are flagged immediately; cumulative changes trigger recommendations
6. Monthly reports are auto-generated and sent via WhatsApp

Your RECIPES.md file is simple — you describe each dish's ingredients and approximate quantities. The AI handles the math.

How We Set This Up

None of this works if the AI is just a standalone chatbot with no connection to your actual business. That's why BestAI builds a custom integration program — a piece of software that bridges your AI assistant with the systems you already use.

For this kind of setup, that means:

• Connecting the AI to WhatsApp and Facebook Messenger so customers can reach you on the channels they already use
• Linking to your accounting or invoicing system so nothing gets lost between quote and invoice
• Setting up email notifications and automated follow-up sequences
• Building your knowledge base from an initial interview and keeping it updated as your business evolves

Here's our process:

1. We map your current workflow — We sit down with you and figure out what tools you're using and how enquiries currently flow through your business.
2. We build the connections — Our developers write a custom program (an API connector) that lets the AI talk to your systems. No manual data entry, no copy-pasting between apps.
3. We test end-to-end — Every workflow gets tested with real scenarios before going live. Nothing launches until it works reliably.
4. We maintain it — When your business changes, we update the integration to match.

You don't need to be technical. We handle all the development — you just tell us how your business runs, and we make the AI fit into that.

The Result

• Real-time cost awareness — know your food cost percentage this week, not last month
• Automatic price change alerts — supplier increases flagged immediately
• Per-dish margin visibility — know which dishes make money and which don't
• Supplier comparison — data-driven decisions about where to buy
• No spreadsheets — invoice photos in, insights out

What AI Can't Do Here

• AI won't negotiate with suppliers — it provides the data, you have the conversation
• AI can't account for waste — actual food cost includes waste, prep loss, and over-portioning, which need manual tracking
• AI won't set your menu prices — it recommends reviews, you decide
• Accuracy depends on your recipes being reasonably accurate — garbage in, garbage out

Who This Is For

• Single-location restaurants and cafes without enterprise food management software
• Owner-operators who set menu prices by instinct rather than data
• Any restaurant where the monthly P&L is a surprise rather than an expectation
• Businesses feeling the squeeze of rising ingredient costs in 2025-2026