The Monday Prawn Problem

The Real Problem

Chen runs a Chinese seafood buffet in East Auckland. His buffet is popular -- packed on Friday and Saturday nights, steady on Sundays. But Monday through Thursday? It's a guessing game.

Last Monday, he prepped 12 kilograms of prawns, 8 kilograms of mussels, and two trays of crab. By 9pm, half the seafood station was untouched. Into the bin. That's $200 to $400 of product thrown away on a single quiet night. It happens two or three times a week.

Then Saturday comes. The prawns run out by 7:15pm. The crab legs are gone before the second sitting. A family of six walks up to an empty seafood station, and by Monday there's a 2-star Google review: "Seafood buffet with no seafood after 7pm. Don't bother."

This is the fundamental tension of running a buffet: prepare too much and you throw money away. Prepare too little and you destroy your reputation. Academic research puts it bluntly -- approximately 50% of food displayed in buffet services is wasted, making AYCE buffets the most food-waste-generating operations in hospitality.

In New Zealand, food waste runs to 1.2 million tonnes per year nationally. Buffet operators absorb a disproportionate share of that cost. The food cost target for a buffet is 32-38% of cover price. Waste pushes it well above 40%, and on a business where margins are already razor-thin, that's the difference between profit and loss.

Look at Valentines, New Zealand's largest buffet chain -- 9 North Island locations, $3.38M revenue, but only 13 employees. That's how lean these operations run. There is no room for waste.

The real killer is unpredictability. A rainy Tuesday and a sunny Tuesday in school holidays are completely different nights. A Warriors game at Mt Smart drains East Auckland of foot traffic. A public holiday Monday triples your usual covers. Chen can't predict any of this on instinct alone, and he doesn't have data to help him.

Why Existing Tools Don't Solve This

Enterprise food waste platforms exist. Winnow uses camera systems and AI to monitor plate waste. Leanpath tracks pre-consumer waste in commercial kitchens. But these are designed for hotel chains and institutional caterers -- Marriott, Hilton, hospital kitchens. They cost thousands per month and require hardware installation.

For a single-location NZ buffet doing $600K-$1.5M in revenue, that's not realistic. Chen's POS (Lightspeed) tracks how many covers he does each night. Xero tracks his food purchases. But nothing tells him "prepare 40% fewer prawns tonight because it's raining and there's no school holiday traffic."

Chen runs on instinct. His instinct is wrong two or three nights a week. That's $400 to $1,200 in wasted product every single week.

How AI Solves This

Every morning at 6am, Chen gets a WhatsApp message with today's prep recommendations:

Buffet Prep Forecast -- Monday 14 April

Expected covers: 62 (confidence: high) Comparison: last Monday 58 covers, last rainy Monday 49 covers

Recommended prep by station:

• Seafood: Reduce 40% from weekend levels. Prawns 7kg (not 12kg). Mussels 5kg (not 8kg). Skip crab legs -- Monday demand doesn't justify cost.
• Roast/carving: Maintain standard. Roast pork + lamb. Monday regulars come for the roasts.
• Stir-fry/hot dishes: Standard prep. Consider adding congee -- high uptake on cold evenings (tonight: 14C forecast).
• Dessert: Increase 10%. Monday crowd skews older, dessert uptake higher than average.

Why this forecast: Rain forecast from 5pm. No school holidays. No major local events. Last 6 Mondays averaged 61 covers. Similar weather Mondays averaged 52.

Chen adjusts his prep accordingly. He trusts the data because he can see the reasoning.

Mid-Session Adjustments

At 6:30pm, 45 minutes into dinner service, a staff member logs current station levels:

Staff: "Prawns 60% remaining, roast pork 30% remaining, dessert station full"

AI: Based on current pace (38 covers served, 24 remaining), prawns will finish the night with approximately 2kg surplus -- within acceptable range. Roast pork may run low by 8pm. Recommend slicing the backup lamb shoulder now.

Weekly Waste Report

Every Monday morning:

Waste Report -- Week of 7-13 April

Total food cost: $8,240 (34.2% of revenue -- target: 35%) Estimated waste: $380 (down from $920 average before AI) Best-predicted night: Saturday (forecast 185, actual 181) Worst-predicted night: Thursday (forecast 90, actual 118 -- school concert nearby not in events data)

Learning note: I've added "Pakuranga College events" to the local events watchlist. This should improve Thursday forecasts.

How It Works Under the Hood

1. AI ingests 6+ months of POS data -- covers by day, by hour, by session (lunch vs dinner)
2. Pulls weather forecasts via API every morning
3. Checks local events (school calendars, sports fixtures, public holidays)
4. Cross-references booking data if online bookings are enabled
5. Generates per-station prep recommendations using learned consumption patterns
6. Staff log leftover quantities after each service -- the AI learns and improves

The system gets smarter every week. After 3 months, it knows that rainy Mondays average 15% fewer covers, that school holidays lift weekday dinner by 25%, and that the seafood station over-indexes on Friday nights.

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 your POS system (Lightspeed, Square, or whatever you use) so it can pull historical cover data automatically
• Integrating weather forecast and local events APIs so the AI has real-time context
• Setting up a simple WhatsApp-based logging system for staff to report station levels and end-of-night waste
• Building the daily forecast delivery via WhatsApp so it fits into your existing morning routine

Here's our process:

1. We map your current workflow -- We sit down with you and figure out what data you already have (POS history, booking records, supplier invoices) and how prep decisions currently get made.
2. We build the connections -- Our developers write a custom program (an API connector) that lets the AI talk to your POS, weather services, and event calendars. No manual data entry, no spreadsheets.
3. We test end-to-end -- We run the system in parallel with your existing prep process for 2-4 weeks, comparing AI forecasts to actual results before you start relying on it.
4. We maintain it -- When you add new stations, change suppliers, or adjust your menu, we update the system to match.

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

The Result

• 40-60% reduction in food waste -- AI demand forecasting studies show 14-52% reduction in wasted meals; real-world buffet operators report even higher savings
• Per-station prep guidance -- not just "expect 65 covers" but "reduce prawns, maintain roasts, add congee"
• Weather and event awareness -- the AI factors in what you can't see from inside the kitchen
• Continuous improvement -- the system learns from every service, getting more accurate over time
• No more Monday prawn problem -- expensive seafood only gets prepped in quantities that match actual demand

What AI Can't Do Here

• AI won't cook the food or manage your kitchen -- it provides data, your chef makes the calls
• AI can't predict one-off surprises -- a tour bus of 40 showing up unannounced is outside the model
• Accuracy depends on staff logging waste honestly -- if nobody reports leftovers, the AI can't learn
• The first 4-6 weeks are a learning period -- forecasts improve significantly after the system has enough data
• AI won't replace good kitchen management -- a great chef with bad data will outperform a mediocre chef with good data

Who This Is For

• Buffet restaurants and AYCE operations losing money to food waste
• Seafood buffets where the high cost of premium ingredients makes waste especially painful
• Any buffet operator who's guessing prep quantities based on instinct
• Restaurants where weekend and weekday demand varies wildly and unpredictably
• Buffet owners who've had bad Google reviews because popular items ran out too early