The Real Problem
Sung-ho runs a 20-table all-you-can-eat Korean BBQ restaurant in Auckland CBD. The concept is simple: $49 per head, eat as much as you want. Customers love it. The dining room is packed every weekend.
But Sung-ho's margins are disappearing.
His food cost target is 30%. Last month it crept to 38%. The month before, 36%. The trend is going the wrong direction, and he knows exactly why -- but he can't fix it with the tools he has.
Here's the problem with AYCE: customers order more than they eat. A table of four orders eight plates of beef, six plates of pork belly, four seafood platters, and a mountain of banchan (complimentary side dishes). Half of it ends up in the waste bin. Academic research identifies AYCE restaurants as "the most food-waste-generating service operations" in the industry.
Sung-ho tried waste penalties -- a $10 charge per plate of uneaten food. It's common in Korean BBQ. But it creates terrible customer friction. His Google reviews started dropping: "Got charged $30 for leftover food, won't come back." One bad review about waste penalties can cost more than the penalty collects.
He tried limiting orders to two items at a time -- a manual process where staff monitor tables and say "you can order more when you finish what's on the grill." This works in theory but breaks down in practice. His staff are also managing grills, clearing plates, restocking the sauce bar, and preparing 6 to 8 banchan dishes per table daily. Nobody has time to watch 20 tables' ordering patterns.
Meat inventory is 50 to 65% of his food costs. When he preps too much brisket for a Tuesday that turns out slow, it goes to waste. When he preps too little for a Saturday, he runs out at 8pm and has to turn customers away. Every day is a guessing game.
Why Existing Tools Don't Solve This
Enterprise food waste platforms exist -- Winnow and Leanpath are the big names. They use smart scales and cameras to track what gets thrown away. But they cost thousands per month and are designed for hotel chains and corporate cafeterias, not a 20-table BBQ restaurant.
Sung-ho's POS tracks orders but not consumption. It knows table 7 ordered five plates of beef. It doesn't know table 7 only ate three. It doesn't know that table 12 always orders the same pattern -- heavy on premium cuts early, then barely touches the last round. It doesn't flag that Thursday prep quantities should be 30% lower than Saturday.
Regular waste audits -- physically weighing what gets thrown away -- can save $2,000+ annually per restaurant. But Sung-ho doesn't have staff to stand by the bin with a scale. He needs insight without overhead.
How AI Solves This
Sung-ho's AI system works behind the scenes, tracking ordering patterns and turning them into actionable data.
Smart Ordering Tracker
When staff enter orders into the tablet POS, the AI tracks consumption per table in real time:
Table 7 -- Alert
4 guests, 55 minutes into session Ordered: 6 plates premium beef, 4 pork belly, 3 seafood Pattern: High volume early, slowing down. Similar tables typically order 1-2 more rounds max. Suggestion: If they order another premium beef round, confirm with a friendly check-in.
This isn't a "stop them from ordering" system. It's information for staff. Instead of a waste penalty, the server can say:
"How's everything going? Still hungry, or can I get you something lighter -- maybe some vegetables for the grill or a cold noodle?"
A gentle redirect instead of a confrontation. The table feels cared for, not policed.
Daily Prep Prediction
Every morning, the AI generates a prep sheet:
Prep Sheet -- Saturday 5 April
Expected covers: 140 (based on reservations + historical Saturday average) Weather: Rain forecast -- historically +12% covers on rainy Saturdays Adjusted expected covers: 157
Recommended prep quantities:
- Beef brisket: 28kg (last Saturday: 25kg, ran short at 8:45pm)
- Pork belly: 18kg
- Premium short rib: 12kg
- Seafood platter components: 35 sets
- Banchan (per table): 6 sides x estimated 42 table turns
Tuesday comparison: Last Tuesday: 65 covers. Prepped for 90. Surplus: 8kg beef, 5kg pork (stored/wasted). This Tuesday recommendation: prep for 70, with 15% buffer.
The AI learns from actual consumption data over time. It knows that rainy weekends bring more diners. It knows Tuesday is slow. It knows school holiday weeks spike by 20%. No more guessing.
Waste Pattern Reports
Weekly, the AI sends Sung-ho a waste summary:
Waste Report -- Week of 31 March
Estimated food waste: 42kg (down from 58kg previous week) Top wasted items: beef brisket (12kg), banchan kimchi (8kg), seafood (6kg)
Patterns:
- Tables ordering 5+ plates of premium beef in the first 20 minutes waste 40% more than those who pace orders
- Banchan waste is highest on weeknights when covers are lower but prep is the same
- Seafood waste spikes on Mondays (quality perception -- customers avoid Monday seafood)
Recommendations:
- Reduce Monday seafood prep by 30%
- Adjust banchan batch sizes: weeknight batches should be 60% of weekend
- Consider removing one banchan option on slow nights to reduce variety waste
Margin Dashboard
Sung-ho checks his phone each morning:
Daily Margin Snapshot -- Friday 4 April
Revenue: $6,370 (130 covers x $49) Estimated food cost: $1,848 (29.0%) Target: 30% -- you're under target today.
Week-to-date: 30.2% food cost (on track)
When the number creeps above target, Sung-ho knows immediately -- not at the end of the month when it's too late.
How We Set This Up
None of this works if the AI is just a standalone dashboard with no connection to Sung-ho's actual restaurant operations. That's why BestAI builds a custom integration program -- a piece of software that bridges the AI system with the tools Sung-ho already uses.
For this kind of setup, that means:
- Connecting the AI to the POS system so every order is tracked automatically -- no manual data entry
- Building the prep prediction engine using historical sales data, reservation counts, and external factors like weather and school holidays
- Setting up real-time alerts that go to staff tablets or Sung-ho's phone when consumption patterns suggest waste risk
- Creating automated daily prep sheets and weekly waste reports delivered via WhatsApp or email
Here's our process:
- We map your current workflow -- We sit down with Sung-ho and understand how ordering, prep, and waste currently work. Every AYCE restaurant runs differently.
- We build the connections -- Our developers write a custom program (an API connector) that lets the AI read POS data, reservation numbers, and generate predictions. No manual tracking, no spreadsheets.
- We test end-to-end -- We run the system alongside normal operations for two weeks, comparing AI predictions against actual results, before trusting it for prep decisions.
- We maintain it -- When Sung-ho changes his menu, pricing, or layout, we update the system to match.
You don't need to be technical. We handle all the development -- you just tell us how your restaurant runs, and we make the AI fit into that.
The Result
- Food cost back under control -- from 38% down toward the 30% target, tracked daily instead of monthly
- Smarter prep quantities -- less surplus on slow nights, fewer shortages on busy ones
- No more waste penalties -- replaced with intelligent portion management and gentle staff check-ins
- Banchan waste reduced -- batch sizes matched to actual covers, not a fixed daily quantity
- Daily visibility -- Sung-ho knows his margin every morning, not just at month-end
- Staff empowered -- servers have data to guide conversations instead of confrontations
What AI Can't Do Here
- AI can't physically stop customers from over-ordering -- it provides data and suggestions, not enforcement
- AI won't weigh waste directly -- estimates are based on ordering patterns vs. historical consumption, not bin-level measurement
- AI can't account for kitchen prep waste (trimming, spoilage) -- only front-of-house ordering waste
- AI predictions improve over time but aren't perfect from day one -- the system needs 4 to 6 weeks of data to calibrate well
- AI won't replace good kitchen management -- a head chef who knows their prep is still essential
Who This Is For
- AYCE Korean BBQ, hotpot, or buffet restaurants struggling with food cost creep
- Restaurants that have tried waste penalties and seen customer backlash
- Owners who prep based on gut feeling and regularly over-prep or under-prep
- Any AYCE venue where food cost is above 33% and trending in the wrong direction
- Restaurants with 15+ tables where staff can't manually monitor every table's consumption