Latest / From Kitchen to C-Suite™ / Line Cook to Code: Beating Margin Squeeze with Agentic AI

Transcript

Line Cook to Code: Beating Margin Squeeze with Agentic AI · From Kitchen to C-Suite™

• Brad Peters: So To the vision that This is Kitchen to see Sweet. Welcome to the show. So I want you to picture a Friday night at your favorite local independent restaurant. You know the vibe, right? yeah, totally. The dining room is packed. There's that dull roar of conversation, glasses clinking, the energy is just electric. Exactly. But if you push through those swinging doors and walk into the kitchen, the reality is, well, it's entirely different. It is basically controlled chaos. Highly controlled chaos. And honestly incredibly fragile. Right. So I want you to imagine a line cook. Okay, they're working the grill station, getting slammed, and they accidentally drop a beautiful, like a forty dollar prime ribeye steak right onto the floor. that hurts just thinking about it. Right. Now in a normal business, he'd just lose the cost of the product. But in the traditional restaurant industry, because those financial margins are so incredibly razor thin, I mean, historically hovering between a staggering three percent and maybe maybe seven percent if the operator is just really lucky? Yeah, seven percent is a great year for a lot of places. Seriously. So that single dropped stake doesn't just mean a loss of inventory. It means that the restaurant effectively makes zero profit for that entire hour of service. Literally zero. Yeah. The entire team is suddenly just, you know, working for free during that hour. Yeah. And it's a phenomenal way to frame the reality of this business because for decades we've kind of I don't know, romanticize the grit of the culinary world. Aaron We really have. Yeah. We focus on the heat, the pressure, the whole, you know, artistry of the plate. But behind that artistry is a financial landscape that has just become absolutely brutal. Brutal is the right word. It is. We're looking at this compounding systemic compression of margins that is basically pushing traditional operators toward absolute zero. Aaron So for you listening right now, whether you're prepping for an operational board meeting. Or you're trying to figure out how to manage an overwhelmed team, or even if you're just insanely curious about how the food on your plate actually gets to your table without, you know, bankrupting the person who cooked it, this deep dive is built for you. Absolutely. Because we are examining a massive transition happening in hospitality right now. Yeah, the old brute force method of running a kitchen where you're relying on paper timers, clipboards, a chef's memory, and just pure Exhausting human grid. It's fundamentally breaking down under modern economic pressures. Aaron Well, yeah, because those pressures are coming from every single side simultaneously. Like you have extreme food cost volatility. Right, like what we saw recently. Exactly. A drought in one part of the world suddenly spikes the cost of cooking oil or an avian flu outbreak just doubles the price of eggs overnight. Which is terrifying if you run a breakfast place. it destroys your budget. Mm-hmm. Then you layer on the third party delivery platforms. They're taking Anywhere from fifteen to thirty percent of the top line ticket size. Thirty percent. I mean that alone wipes out a traditional margin. It totally does. And then you have rising energy costs, skyrocketing commercial rent, and an entry level workforce that is turning over at a rate exceeding seventy-five to a hundred percent annually. The math of the traditional restaurant is just broken. It's completely broken. Which really brings us to the mission of our deep dive today. We are exploring the shift from what we might call, you know, line cook operational instincts. To C-suite software-driven execution. And we've got some amazing material to pull from for this. We really do. To do this, we've assembled this phenomenal stack of sources. We're synthesizing operational blueprints from modern hospitality groups, some deeply technical case studies on yield management, including frameworks from SAS, and real-world deployment data from computer vision platforms like Lean Path and Winnow. Plus, we're digging into the human element, which is huge. We're looking at workforce strategies designed to end those toxic kitchen cultures. Yeah. And surprisingly, we're pulling insights on transitioning into tech from the platform SheCanCode. Right. Because the core mechanism we're analyzing across all these sources is the deployment of a very specific cutting edge technology known as agentic AI. Agentic AI. Yeah. Not just regular AI. Exactly. Operators are using this not just as some flashy novelty, but as an actual survival mechanism to automate back of house operations. The goal here is to aggressively defend what the industry calls GPAR, right? Yeah, GPAR. Gross operating profit per available room or seat. And they want to completely rebuild the financial engine of the kitchen without, you know, stripping away its culinary soul. I want to pause on that acronym for a second, G-O-P-P P A R, because it really sets the stage for everything we're about to discuss today. Gross operating profit per available seat. But it's such a crucial metric. It really is. Because if you think about a restaurant seat, it's essentially real estate. The operator pays rent, utilities, and insurance for that specific square footage twenty four hours a day, seven days a week. Regardless of whether anyone is sitting in it. Right. So if that seat is empty on a Tuesday at three PM, they are actively losing money on it. And if someone sits in it, orders a five dollar side of fries and just I don't know, camps out for three hours, they are also actively losing money on it. Yeah, the math doesn't lie. Maximizing the profit generated by that specific physical space is basically the entire game. And doing that requires a level of forecasting and analysis that a human being simply cannot do in real time, especially while they're also, you know, trying to actually cook the food. No, it's impossible. And the distinction we really need to make right out of the gate here is what we mean when we say AI. Because let's be honest, artificial intelligence is this massive, somewhat bloated buzzword right now. totally. Everything has AI slapped on it. Right. A lot of platforms claim to have AI when they really just have a slightly advanced sorting algorithm. So we are strictly looking at practical, margin saving, autonomous tools that catch capital leakage before it ever hits the monthly PL statement. I've always thought of traditional restaurant management as trying to drive a speeding car down this winding mountain road, but you're only looking in the rearview mirror. That is a terrifying image, but very accurate. It's true. You print out last week's sales averages, you look at last month's inventory counts. And you try to steer today's chaotic Friday night shift based entirely on what happened in the past. Which is just a recipe for disaster in such a volatile market. Exactly. So what we're exploring today is how this agentic AI acts as a GPS that actually looks at the road ahead, analyzes the traffic, senses the weather, and then, you know, actually turns the steering wheel for you. I love that analogy. Thanks. So let's start with the macro squeeze. How bad is the financial erosion that is forcing this whole technological shift? Well, to truly understand the erosion, we have to look at the foundational math of the restaurant business, which is the prime cost formula. Okay, prime cost. Break that down for us. So your prime cost is made up of two things. First, your costs of goods sold, so the physical ingredients you buy, like the proteins, the produce, the packaging. Right. The physical stuff. Yeah. And then plus your labor costs, which includes payroll, taxes, and benefits. So in a standard traditional casual dining or quick service model, A healthy prime cost is supposed to absorb about sixty-four percent of the net revenue. Okay, so out of every dollar a customer spends, sixty-four cents goes immediately just to buying the food and paying the people to cook it and serve it. Precisely. That's sixty-four cents gone right away. Then you factor in occupancy and operational expenses rent, electricity, marketing, insurance, repairs, which typically eat up another twenty nine to thirty percent. Wow. So that leaves you with what, at best a four to six percent net margin? At best. Yeah, a four to six percent net margin. And the crisis happening right now is that the sixty four percent prime cost is creeping up. Because of the inflation we talked about. Exactly. Food costs fluctuate wildly and labor wages are rising. So when your prime costs hit sixty eight or seventy percent, that tiny remaining four percent margin is just entirely obliterated. It's gone. Poof. The restaurant is operating at a deficit, even if the dining room is completely full, profitability just collapses. That is so wild to think about. A packed restaurant that's actually losing money. But this is where the new paradigm enters the picture, right? The agentic model we found in our sources is entirely focused on driving those prime costs down from that dangerous mid sixties percentage to around fifty-one percent. Which is a massive shift. It's huge. If you can use technology to get your cost of goods sold down to twenty-six percent. And your labor optimized to 25%, suddenly you've unlocked a massive 17% net margin. Yeah. And 17% is a completely different business model. Aaron It turns a fragile mom and pop operation into this highly lucrative, scalable enterprise. But the real question here is how on earth do you shave 13% off your prime costs without, I don't know, serving lower quality food or just firing half your staff? That's the magic question. And the answer lies in the transition to agentic AI. And we really need to clarify what makes a technology agentic. Right, because we said it's different from standard tech. Very different. For the last decade, restaurants have adopted passive technology. Like a digital ordering kiosk in the lobby is passive. It just waits for a customer to push a button. Or like those screens in the kitchen. Yeah. A kitchen display system or KDS that puts digital tickets on a screen instead of printing them on paper, that is passive. It just displays information. Passive technology digitizes existing bottlenecks, but it doesn't actually solve them. So what does Agentik do differently? Agentic AI is an autonomous software workflow. Meaning it has agency. It acts on its own. Yes. It perceives environmental inputs, it makes complex, multi-step decisions based on those inputs, and it takes physical or digital actions in the real world without a human manager ever having to click approve. I'm struggling a bit to visualize what that Actually looks like in a kitchen. Can you give me a concrete scenario? Like what does agentic AI do on a Tuesday morning prep shift that a human sous chef doesn't do? Sure. Let's look at dynamic par ordering, which is basically how a restaurant decides what groceries to buy for the week. Okay, par ordering. Yeah. So in a legacy system, the sous chef walks into the walk-in refrigerator with a clipboard. They look at a shelf, see they have, say, 10 pounds of salmon left. They know from memory that they usually sell 50 portions on a Tuesday, so they write down an order for 40 pounds of salmon, go to a computer in the office, and email the supplier. Aaron So it's entirely based on their static memory and just a visual guess of what's on the shelf. Right. And humans miscalculate that all the time. Now, an agentic AI system completely removes the clipboard and the guessing. The AI is constantly reading real time item velocity from the point of sale system. Aaron So it knows exactly what's selling right now. Down to the minute. It knows exactly how fast the salmon is selling, but it doesn't stop there. It actively pulls in the hyperlocal weather forecast. Wow. Okay. It scans the upcoming reservation books from platforms like Open Table or Resi to see the exact party sizes coming in. And it monitors live supplier price sheets. So it's absorbing millions of data points across all these different categories that a human couldn't possibly cross-reference in their head at one time. Impossible. And here is the agentic part. Let's say the system sees a massive storm front approaching in the weather data for Thursday. Okay, so storm's coming. It queries its own historical database and determines that historically, heavy rain depresses dinon seafood sales by twenty two percent, while simultaneously spiking delivery orders for comfort food like pasta. Because nobody wants to go out in the rain to eat fish, they want to stay home and eat spaghetti. Exactly. Human nature. So the AI doesn't just send the chef a little push notification saying, Hey, heads up, it might rain. It autonomously reduces the raw seafood order for Wednesday's delivery by 22%. On its own. On its own. It increases the dry pasta order, and it issues the purchase orders directly to the suppliers via vendor APIs. Okay. I want to break down that API concept for the listener real quick because it's such a critical piece of the plumbing here. An API or application programming interface is essentially a secure digital bridge. That allows two completely different software systems to talk to each other, right? Yeah, that's a perfect way to put it. It's like two computers handing each other a digital clipboard in a language they both instantly understand. So the restaurant's AI talks directly to the seafood supplier's inventory computer, places the order, locks in the bulk pricing, and updates the accounting ledger, all in a fraction of a second, without a single human ever typing an email. And it does this continuously. Over and over. Optimizing for the lowest possible cost and the exact right amount of inventory, it eliminates the massive margins of error that come with human guesswork. I do have to challenge this though. I mean I'm putting myself in the shoes of a veteran chef who has spent, you know, twenty years honing their craft, learning how to read their dining room, learning how to select the best ingredients. Right, the artistic side of it. Yeah. If you have a piece of software autonomously deciding what fish to buy, what vegetables to bring in, doesn't that strip the intuition out of the kitchen? Doesn't it turn a culinary artist into just like an assembly line worker for a machine? What happens when the algorithm inevitably gets it wrong and under orders for a massive dinner rush? Aaron It's the most common and honestly the most valid fear in the industry right now. But it fundamentally misunderstands what the AI is actually replacing. Aaron Okay, how so? The AI is not replacing the chef's palate. It is not replacing their ability to conceptualize a new dish or build a flavor profile or Mentor a young cook on how to properly sear a scallop. Right. It's not cooking the food. Exactly. The AI is replacing the frantic, soul-crushing mental math that a human being is forced to do under extreme stress. so it's taking away the busy work. It's taking away the cognitive overload. Look, the human brain is incredible, but a human manager cannot possibly hold five moving variables in their head at the same time. Like the weather and the prices. Right. The bare metric. Pressure, the local concert schedule, the live reservation cancellations, the historical sales velocity of 50 different menu items, and the current stock levels in the walk-in. You can't hold all that in your head simultaneously while also expediting dinner service for 200 angry guests. Yeah, that's just a recipe for a breakdown. When humans try to do that, they fail. They overorder, and that failure results in a 15 to 25% ingredient spoilage rate. Food literally rots in the fridge because the chef guessed wrong. Which kills that tiny margin. Right. So by catching what human memory cannot, the AI causes a direct one and a half to three percent drop in the cost of goods sold. That makes a lot of sense. You're shifting the labor away from mental calculation and redirecting it toward tactical, high quality execution. You are essentially telling the chef, Hey, stop being a human calculator and go back to being a chef. Make the food taste incredible. Let the machine do the algebra. Exactly. If the AI is standardizing the purchase yields and ensuring the exact right amount of food is in the building, the culinary team is freed from that administrative panic. They can focus purely on the hospitality and the execution of the dish. Okay, so this transition from buying the food to pricing the food is where the strategies become just aggressively modern. Because optimizing what comes into the kitchen is really only half the battle. absolutely. You have to optimize what you charge for it on the way out. Our sources dive deeply into revenue management and surgical menu engineering. And they draw this fascinating parallel between the hospitality industry and the way airlines or hotels price their inventory. Because the underlying mathematics are totally identical. Really? A restaurant and an airline. Yeah. Think about it. Any business that has fixed and limited capacity, perishable inventory, customers who can be segmented by their willingness to pay. And demand that varies wildly by the hour or the day faces the exact same optimization puzzle. Perishable inventory is definitely the key phrase there. I mean, we understand that a tomato is perishable, it rots. But a seat in a restaurant is also perishable in a temporal sense, isn't it? Completely. If a four top table sits empty from six PM to seven PM on a Friday night, the restaurant can never, ever sell that specific hour of that specific table again. The potential revenue from that time slot is basically spoiled and gone into the garbage just like a rotten tomato. The SAS revenue optimization framework we reviewed breaks down how to manage this temporal perishability beautifully. I loved that section. That's so good. They use the classic hotel revenue management model to illustrate the point. They detailed a hypothetical story of Jane, a revenue manager for a seven hundred room City Center hotel, who is trying to set pricing for Memorial Day weekend. Let's walk through Jane's dilemma because The logic maps perfectly to a restaurant dining room. So Memorial Day is a massive weekend. Most travelers want to arrive on Friday and check out on Monday for a full three night stay. Right. But Jane also has customers calling who just want a room for Saturday night, which is the absolute peak of the weekend. The busiest night. Yeah. And this creates what economists call a network effect problem. If Jane sells a room to the guest who only wants it for Saturday, she physically blocks that exact room from being sold to a guest who might have wanted it for Friday, Saturday, and Sunday. So the sale of a one night product inherently cannibalizes the sale of a three night product. Exactly. So mathematically, Jane has to realize that even if she charges a premium for the one night, say she bumps it to two hundred and fifty dollars just for Saturday, she is actually losing money compared to selling three nights at a discounted rate of one hundred and fifty dollars a night. Because three nights at one fifty is four hundred and fifty dollars total. Right. She'd be throwing away two hundred dollars by being short sighted. But the complexity there is that Jane doesn't know for sure if the three night guest will actually book, right? Like that booking hasn't happened yet. So how does she make that decision? Well, she relies on algorithms that calculate what the industry calls unconstrained demand. Unconstrained demand. What does that mean in practice? This means the software analyzes historical data to determine how many rooms they could sell if they had an infinite number of rooms at all possible price points. wow. So it calculates the theoretical maximum demand. Yes, the absolute theoretical maximum. And based on that, the AI recommends the optimal mix of restrictions. It tells Jane, hey, do not sell any rooms for just Saturday night until we are 90% full, because the data says the three night bookings will eventually materialize. And restaurants are finally adopting this exact logic. Yes. Because a table of two that sits down at 7 p.m. and orders only appetizers and tap water. Is blocking a table of two that might have ordered a four-course meal and two bottles of wine. That is fascinating. And identifying the profitability of what the guests actually order brings us to the Masteront case study, which completely blew my mind. It's a game changer. This is the application of algorithmic pricing to the physical menu itself. The case study looks at an operator with a 38-dish menu, and they took the standard written recipes, the real purchase prices of the ingredients from their invoices. And the actual sales mix, meaning exactly how many of each dish were actually sold from their point of sale system. Right. And they fed all of this into an AI costing engine. Aaron Because the goal was to pit the theoretical cost of the dish against the real cost, dish by dish, dynamically over time. Aaron So theoretical cost is what the dish should cost if every cook measures every single ingredient perfectly according to the recipe book. Which never happens. Never. And the real cost is what it actually costs. based on how much bulk food was purchased versus how much revenue was ultimately generated. Right, capturing all the waste and heavy handed pouring. So the AI computed each dish's true contribution margin, ranking the entire thirty-eight item menu by its EBITDA impact rather than just its overall popularity. I want to define EBITDA really quickly for the listener, because it's a term thrown around in business constantly, but it can be confusing. It stands for earnings before interest taxes depreciation An amortization. A mouthful. Right. But in plain English, especially in a restaurant context, EBITDA is the actual physical cash you have left in your hand from selling the food before the bank takes its loan payment and the government takes its taxes. It is the true measure of your operational cash flow. So the AI ranked the menu by which dishes actually put cash in the bank. And the revelation from the AI's analysis was just staggering for this operator. Out of thirty eight dishes, only six, just six, were concentrating almost all of the positive EBITDA margin. Wow. The entire business was surviving on the backs of six items. Meanwhile, five of their quote unquote heavy sellers, dishes that the customers ordered constantly, dishes the chef was incredibly proud of, they were actually bleeding money. Wait, how does a top selling dish bleed money? If you're selling a ton of shouldn't you be making money? You'd think so. But it happens because the real food cost was measuring several points higher than the theoretical cost. Give me an example of how that happens on the line. Sure. Let's say it's a braised short rib dish. Theoretically it's supposed to cost 30% of the menu price to make. But in reality, the supplier has been slowly raising the price of beef over the last three months, just a few cents at a time. And the line cooks are being heavy handed with the expensive wine reduction sauce when they plate so a little extra sauce here. A slight price hike there. Exactly. The real cost pushes to forty percent instead of thirty. So every time a customer orders it, the restaurant is actually losing margin compared to their baseline. Aaron That gap between the theoretical thirty percent and the actual forty percent is literal cash just evaporating into thin air. It really is. It's invisible waste, unstandardized portioning and supply chain inflation that a human manager, who's maybe reviewing spreadsheets once a month Simply cannot catch in time. The AI didn't rely on gut feeling. It put hard, undeniable mathematical numbers on the board. But here's the thing. Diagnosing the disease is very different than curing That's true. The operator's initial gut instinct when faced with this terrible data was to freeze. They were terrified to adjust the prices on those five top selling money losing items. Why? Because they believe that if they raise the price by even a dollar, their loyal customers would revolt and just stop coming in entirely. Which is a completely understandable human fear. You don't want to bite the hand that feeds you. But you also cannot run a business where your most popular products are basically subsidizing the customer's meal. No. You'll go bankrupt. Right. So how does the AI solve the pricing problem without triggering a massive customer boycott? This is where the genius of AI driven elasticity pricing comes in. You don't just blindly mark up the entire menu by 10% across the board, which is what panicked operators traditionally do when they realize they're losing money. Yeah, that usually ends badly. The AI analyzes historical elasticity data. It looks at millions of past transactions to see how demand for a specific item changed when the price fluctuated in the past. It scrapes competitor pricing from the web within a five mile radius. It correlates items that are frequently bought together. So it finds the Hidden patterns in consumer behavior. Exactly. The AI identifies which specific items have inelastic demand. Inelastic means the customer is insensitive to price changes. Aaron Okay, so they don't care if the price goes up a bit. Right. They are coming to your restaurant for that specific signature dish, and they will pay an extra two dollars for it without even blinking. Conversely, the AI identifies which items are value anchors. Value anchors. Yeah. Dishes like a basic burger or a side salad where even a fifty cent increase will cause a massive drop in orders because the customer perceives it as a ripoff. So it's surgical. In this case study, they didn't touch the value anchors at all. They selectively raised the check average only on those highly inelastic signature dishes where the data proved the customer would absorb the cost happily. Yep. And for the items where the price couldn't be raised, they redesigned the portion sizes, or they swapped out an expensive garnish to bring the cost down internally. Exactly. That difference between a surgical, algorithmically guided price increase and a blind, panic-driven menu markup is the exact difference between recovering your profit margin and alienating your entire customer base. Aaron And this capability extends far beyond the physical printed menu in the dining room, right? Think about the digital landscape, specifically dynamic delivery pricing. this is huge. Because the wholesale food market is incredibly volatile. If the USDA releases a report stating that the domestic cattle herd is at a seventy five year low, the wholesale price of beef can spike eighteen percent overnight. Literally overnight. And if a human is managing the restaurant, they won't actually feel the pain of that beef spike until the end of the month when the accountant does the books. By which point they've lost thousands. But an agentic AI system connected to the supplier APIs detects that cost spike instantly. Instantly. It can then autonomously modify the restaurant's pricing on third-party apps like DoorDash or Uber Eats to protect the margin on that specific day. Or if the operator doesn't want to raise prices, the AI can temporarily shift the visual hierarchy of the menu on a digital self-serve kiosk or the delivery app. Wait, it changes what the customer sees first? Yes. It can autonomously push the expensive beef items down to the bottom of the screen and bring high margin items like A signature pasta dish or maybe a chicken sandwich right up to the top as a quote unquote chef's recommendation. wow. So it subtly manipulates demand toward the profitable items without anyone noticing. Precisely. And it goes even further during a service rush. Let's say it's seven thirty PM on a Friday, the dining room is packed, the ticket printer in the kitchen is just spitting out orders to the floor, and the line cooks are completely drowning. Plastic Friday night in the weeds. Traditionally, the expeditor is screaming at the manager to run to the back office and manually turn off the tablets for Uber Eats and DoorDash so the kitchen can dig out. Which costs the restaurant potential revenue and really frustrates the digital customers who just want to order dinner. Right. But an agentic AI monitors the ticket times in the kitchen display system. When it sees that the average time to produce a dish has crept from, say, twelve minutes up to twenty two minutes, it recognizes the kitchen is overwhelmed. Aaron Okay, so what does it do? It can autonomously throttle the delivery channels. It might dynamically raise the markup on delivery orders by fifteen percent for the next hour to artificially slow down demand. So fewer people order, but the ones who do are paying a premium. Exactly. It maximizes profit on the few orders that do come through while automatically managing the flow of traffic so the kitchen can recover. I am fascinated by this, but I have a major reservation here. I remember vividly reading in our sources about a massive national US burger chain that faced an absolute public relations nightmare because they announced they were testing surge style pricing, kind of similar to Uber. I remember that. It was a disaster. The public backlash was immediate and vicious. People felt like the corporation was exploiting them just because they wanted a burger at noon instead of two PM. If a restaurant is using AI to dynamically change prices on delivery apps or kiosks, how do they avoid making their loyal Regular customers feel completely price gouged. It is the single most critical tightrope to walk in modern revenue management. Public acceptance of dynamic pressing relies entirely on perceived fairness. Perceived fairness. The psychological framing is everything. If a customer believes they are being penalized for eating at a normal time, they will rebel. The burger chain made the mistake of framing it as charging more when it's busy. So what is the correct algorithmic approach then? You flip the psychological model completely. It's about yield management through incentives, not punishment. Okay. How does that look? You establish a standard price that is highly profitable. You do not arbitrarily raise that standard dine-in price during the Friday night rush. That breaks trust. Right. Instead, you use the AI to offer targeted off-peak discounts. You push a notification to your loyalty app users offering 20% off a specific high-margin combo. If they order on a slow Tuesday afternoon between 2 p.m. and 4 p.m. So you condition the price-sensitive customer to move their behavior to your slow periods, while the price-insensitive customer just pays the full profitable premium during the rush without ever realizing they're subsidizing the business. Precisely. You also use the AI to suggest highly profitable add-ons dynamically based on the customer's cart history. And regarding the delivery platforms, you adjust the markups specifically on the apps to offset the exorbitant commission fees those platforms charge. Which protects your margin without punishing the guest who physically walks through your front door to eat. Exactly. You are optimizing the revenue without violating the social contract of hospitality. It's brilliant. It's invisible optimization. Now, before we dive into how this technology is physically changing the layout of the kitchen and the way we track waste. I want take a moment to acknowledge the networks and resources that make deep dives like this possible. Because navigating this transition from a traditional model to an algorithmic one requires immense support. This podcast is supported by the generous commitment of our sponsors, and we extend a special thank you to HRB Universal. Their dedication to providing comprehensive resources and promoting excellence is key to empowering professionals across the culinary and hospitality industries. We also want to acknowledge the Plate and Poor Collective Equity Partner Program. This initiative exemplifies a true commitment to fostering shared success and equitable growth within the food and beverage sector, helping entrepreneurs build a more inclusive and thriving future. The journey to the C-suite is fueled by innovation, relentless dedication, and the right network. We encourage you to engage with the resources and support offered by our sponsors. For details on how to connect with HRB Universal and the Plate and Poor Collective Equity Partner Program, Visit HRBUNI.com or just Google HRB Universal. Okay, let's get back into the kitchen. We have unpacked how AI optimizes the purchasing of the food and how it surgically engineers the pricing of the food. But there is a glaring physical reality we have to confront here. The waste. Yes. How do we stop the actual physical cash from evaporating into the kitchen trash can? I know. That is the million dollar question. The waste problem is the silent killer of the restaurant industry. It really is. The fundamental flaw is that traditional manual inventory counts are entirely retrospective. They happen after the fact. Exactly. By the time a kitchen manager stays late on a Sunday night, physically counts the boxes and the walk in with a clipboard, and realizes that a massive amount of expensive produce spoiled over the weekend, the financial loss is already locked in. You cannot unspoil the food. No. The data is Dead on arrival. The numbers in our sources are genuinely shocking. The industry norm is a four to twelve percent discard rate of all purchased goods. Which is insane. When your net margins is only three to seven percent, throwing twelve percent of everything you buy directly into a landfill is a mathematical death sentence. Which is exactly why the introduction of computer vision waste auditing is perhaps the most revolutionary physical change happening in the back of house right now. Computer vision waste auditing. Yes, companies like Winnow and Lean Path are not just providing software, they are installing highly advanced hardware. They are placing smart scales equipped with high-resolution computer vision cameras directly above the kitchen trash receptacles. I want to visualize this. A kitchen is a chaotic, greasy, fast-paced environment. How does a camera watching the garbage actually function practically during a dinner rush without slowing the cooks down? The beauty of the system is its flictionlessness. The AI is trained on millions of images of food in various states of preparation and decay. So it knows what it's looking at. Yes. Every time a prep cook trims a case of bell peppers, or a dishwasher scrapes a returned half-eaten plate into the bin, or a sous chef throws away an unsold burnt batch of soup at the end of the night, they dump it into a bin sitting on this smart scale. And the camera is watching from above. In a fraction of a second, the computer vision algorithm automatically identifies what the item is. It distinguishes between carrot peels. Burned steak or expired dairy. The scale records the exact weight. Wow. And then the system cross-references that weight with the live API pricing data from the supplier, categorizes the type of waste, and monetizes the financial loss instantly on a digital dashboard. So instead of a manager finding out on Sunday that they lost $100 in spoiled spinach, the chef can look at a screen at 2 p.m. on a Tuesday and see, wow, we have thrown away $42 worth of usable celery trim just today. Exactly. It logs it autonomously without requiring a busy prep cook to pull out a clipboard, wash their hands, and manually write down two pounds of tomatoes. That's incredible. And the results detailed in the case studies really prove the efficacy of this telemetry. IKEA deployed these computer vision scales across their massive restaurant operations, and they cut their food waste by fifty-four percent. Fifty-four percent reduction is massive at that scale. It's unbelievable. And the Conrad Shanghai Hotel. They used this system to save 17,000 meals from going to the landfill, which reduced their overall waste footprint by over forty five percent. And broader inventory management platforms like Total Control reported that digitizing this process resulted in an 80% time reduction for managers who previously spent hours just wrangling inventory operations. The financial ROI is undeniable. But I want to pivot to the human element of this, because our sources from Lean Path highlighted something. Deeply insightful about the psychology of kitchen culture. Yes. The psychology is so important here. You are taking an environment that is historically paranoid and high stress, and you are installing a literal camera over the trash can to watch their mistakes. I imagine the pushback from the staff must be intense. the psychology of implementation is arguably way more difficult than the technology itself. If a corporate director walks into a kitchen and tells a proud executive chef that they are installing a camera to monitor their quote unquote waste, the chef will immediately become defensive. Of course. Because in the culinary world, waste is viewed as a direct accusation of poor technical performance. It implies they don't know how to cook or manage their station. Words matter immensely in change management. So how do you frame it so it doesn't feel like Big Brother is just auditing their failures? Lean Path specifically advises corporate teams to completely eliminate the word waste from the initial rollout vocabulary. Just don't say it. Right. Instead, you frame the technology around improving production compliance or tracking over production. that's clever. A veteran chef inherently understands that sometimes, despite your best estimates, you just prep too much food for a dinner rush that never materializes because of, say, a sudden rainstorm. So by calling it over production. You remove the moral failing. It's no longer about them being a bad cook. It becomes this operational puzzle to solve together. You are giving them a tool to fine tune their prep list rather than a camera to catch them making mistakes. Exactly. And this ties directly into the strategic recommendation to identify a chef champion and a co champion within the kitchen hierarchy. Meaning you need buy in from the kitchen staff themselves. Yes. You cannot mandate this top down from a corporate office, you have to empower the on-the-ground culinary leaders. You emphasize that you trust their ability to run the operation, and this AI tool is simply a magnifying glass to help them find the microscopic, hard to identify inefficiencies that human eyes just naturally miss. That is brilliant psychological engineering. Okay, so we have an AI camera meticulously auditing everything that goes into the trash can. But what about the other side of the physical equation? What about The raw goods coming in the back door from the delivery trucks. Let's talk about invoice auditing. This is where optical character recognition, or OCR, combined with generative AI, is completely overhauling the account's payable process. Aaron Okay, break that down for us. Platforms like MarginEd, which utilizes an AI assistant affectionately named Tom, or systems like Soupy, they are catching the margin leaks that occur right at the loading dock. To appreciate how big of a deal this is, we have to look at the traditional workflow, right? Yeah, paint the picture. A delivery driver rolls a hand truck of produce through the back door. The chef signs a crumpled coffee stained paper invoice, tosses it into a physical tray in the office. And eventually, maybe two weeks later, an exhausted accountant types those line items into an Excel spreadsheet. Which is so inefficient. And if the supplier quietly raised the price of chicken breasts by 14% on that specific invoice, The restaurant will continue prepping, cooking, and selling their chicken dishes for the next two weeks at the old menu price. They won't realize they are actively losing money on every single chicken sandwich until the end-of-month financial review. The delay is literally deadly for the business. So how does the AI invoice auditing compress that timeline? Well, with these new tools, the moment the delivery driver hands over the paper receipt, the chef takes out a tablet or a smartphone and snaps a photo of it right there on the loading dock. Just a quick picture. Yeah. The OCR algorithm instantly reads the crumpled text. The AI extracts every single line item, the quantity, and the price. It digitizes the paper immediately. More importantly, it cross-references that digitized data against the contracted baseline prices in the system's memory. this is where it gets good. If the AI detects that the chicken breast jumped 14% on Tuesday's delivery. It doesn't just log the new price. It instantly pushes an alert to the chef's phone and automatically recosts the digital recipe card for the chicken sandwich before that dish is ever prepped or sold. It flags the price creep in real time. That is amazing. Yes. It catches unauthorized substitutions, it catches missing credits for returned items, and it recalculates the real-time ebita of the menu before the accounting department ever even processes the payment to the supplier. I love this imagery. It is essentially like having a highly aggressive forensic accountant sitting on top of your kitchen trash can watching every scrap of food and another forensic accountant standing on your loading dock auditing every piece of paper in real time. It's exactly like that. It takes the invisible, slow moving margin leaks that destroy businesses and makes them hypervisible and actionable in the present moment. It creates total ground truth telemetry for the physical movement of capital through the building. But and this is a big butt. All of this telemetry, the dynamic pricing algorithms, the smart scales, the automated vendor ordering, it is completely, utterly useless if there is no one left to physically stand over the stove and work the hotline. Because technology doesn't sear a stake. No, it doesn't. Which forces us to pivot to a secondary but arguably the most vital discussion of this entire deep dive. How do we actually use this technology to rebuild the kitchen into a human centric model? We have to address the labor crisis. We just have to, because it is an existential threat to the industry right now. It really is. For over a century, the culinary world has treated frontline labor as this endlessly renewable, highly expendable commodity. The industry was built on the foundation of the Escoffier Brigade system, which is inherently militaristic. It was literally designed to mimic the hierarchy of an army. Exactly. And that resulted in a culture that glorified suffering. It normalized grueling 14 hours split shifts, screaming, public shaming, lack of benefits, and the infamous clopening. man, the clopening. Mm-hmm. Where a cook is scheduled to close the restaurant at 2 AM, clean the fryers, and then return at 8 A.M. to open the restaurant for the next day's prep. It's brutal. Pain is a badge of honor, and if you can't take the heat, get out of the kitchen. That was the explicit mentality. But the demographic reality is just crashing down on that philosophy now. Generation Z will make up the vast majority of the entry-level culinary workforce through the remainder of the 2020s. And they are looking at that bargain, you know, low pay, high abuse, terrible hours, and they are completely rejecting it. They're just walking away to other industries. The result is the staggering turnover rate we mentioned earlier, sitting between 75 and 100% annually. You cannot sustain operational excellence when your entire staff turns over every year. No, and the hard financial impact of that is devastating. The cost to replace a single line cook is estimated to be between $5,800 and $8,200. Wait, I have to stop you there. Where does that massive number come from? A line cook doesn't even make that much money. It's the hidden drag on the business. It's the recruitment, advertising fees, the management time spent interviewing, the cost of training a new person who operates at half speed for the first month. the lost productivity. Right. It's the lost kitchen velocity during a rush because the new cook doesn't know the station. And the cost of the inevitable ticket errors and returned food while they learn. You cannot run a three percent margin business while hemorrhaging eight thousand dollars in invisible costs every time a cook walks out the back door and quits. So how does algorithmic technology, which seems so cold and mathematical on the surface, actually improve the lived experience of the staff and stop this bleeding? Well, it starts with the most fundamental friction point in any hourly worker's life the schedule. The dreaded schedule. Platforms like Forth are using AI demand forecasting to build optimized, humane schedules. Instead of a stressed manager sitting in a back office on a Thursday night, haphazardly copy-pasting last week's shift template, the AI takes over. Because last week's template is a terrible predictor of next week's reality. Precisely. A static template might grossly overstaff the kitchen on a slow Tuesday, which blows up the labor budget. Or it might grossly understaff the kitchen on a Friday night because a local concert was announced. Which leads to total staff burnout and a walkout. So what does the AI do? The AI processes millions of data points to predict the exact labor need, down to 15-minute increments. It analyzes historical sales, local promotional calendars, school holidays, and live weather patterns. Wow, it builds a precision matrix of exactly how many hands are needed on deck at any given moment. Yes. And it cross-references that demand against the staff's input and availability, their preferred maximum hours, and critically, It automatically flags and prevents violations of local labor laws regarding mandatory rest periods and overtime. That's a lifesaver for compliance. It is. It can automatically generate stable four-day block schedules, giving the cooks predictable lives outside of work while simultaneously protecting the restaurant's profit margin. Aaron Predictable hours are a massive step forward. But you know, schedule optimization only goes so far if the job itself feels like a dead end. That's a great point. No one wants to chop onions for exactly 35 predictable hours a week for 10 years without a path upward. This is where our sources introduce the concept of the structural career ladder, and it is a fascinating reimagining of how a restaurant operates. It starts with a dish pit. The dish pit is the absolute heartbeat of any restaurant. If there are no clean plates, no clean saute pans, the entire multi-million dollar operation grinds to a halt in about 20 minutes. Yet historically it is the lowest paid, most disrespected dead end job in the building. Which is such a shame. To retain talent and build a sustainable workforce, modern operators are dismantling the opaque political promotion structures of the past. They are turning these entry level roles into highly structured, transparent apprenticeships. I love this part. They are breaking the entire kitchen operation down into a formalized five level micro credential system. Let's walk through this ladder because this is really the blueprint for the future of culinary management. So level one is the sanitation and logistics specialist, what we used to dismissively call the dishwasher. Mastery at this level is not just about scrapping plates. It requires demonstrating an understanding of chemical safety protocols, were washing throughput optimization, waste sorting compliance, and basic knife care. It's a real job with real skills. Once they master that. They don't just hope for a raise, they officially move to level two, prep apprentice and cold station, which is traditionally known as guardmange. This involves demonstrating mastery of specific knife cuts, emulsifying foundational sauces, building complex vinaigrette, and achieving their basic sanitation certification. From there, they advance to level three, which is the hotline cook. This is where they actually take control of the fry, saute, or grill stations. Mastery here is highly technical. It's about precise temperature control of proteins. station timing, multi-ticket pacing during a rush, and strict allergy management protocols. Level four is the expediter in station lead. This requires advanced cognitive skills. The expediter doesn't just cook, they manage the flow of the entire kitchen. They are responsible for ticket sequencing, ensuring plating consistency across all stations, and critical conflict mediation on the line when tensions run high. And finally level five is the junior sous chef or culinary operations manager. At this level, the physical cooking is secondary. Mastery involves understanding cogs, execute and yield management, interacting fluently with the AI scheduling systems, and handling supplier procurement APIs. What I find so incredibly powerful about this five-level system is how it completely removes the toxic politics from the kitchen. In a legacy restaurant, you get a 25 cent raise if the executive chef happens to like you, or if you wait around for an arbitrary, subjective annual review. It breeds so much resentment. absolutely. But under this micro credential system, every single station features an explicit, transparent, non-subjective checklist of technical skills. He gamifies the progression in the best possible way. If a level two cook wants a raise, they don't have to beg. They request an audit. An audit. Yeah. If they pass the technical checklist over three consecutive audit shifts, it triggers an immediate automated micro raise in the payroll system, plus seventy five cents or Plus $1.25 an hour instantly. It replaces the murky, abusive pay your dues and let me scream at you mentality with a skill-based, undeniable passport to higher wages. And when you combine this with cross-training initiatives, for instance, intentionally paying a level one dishwasher their full wage to shadow a level three line cook for two hours during a slow Tuesday afternoon, you actively dismantle the silos of the brigade system and build deep psychological safety. You are building a pipeline of talent that understands every mechanism of the business, from the chemical mix in the dishwasher to the algorithmic pricing on the menu. Exactly. Which brings us to the ultimate actionable takeaway for you, the listener. If you are going to survive and thrive in this new era of hospitality, whether you are a line cook, a sous chef, or an owner operator, you must fundamentally evolve your skill set. The kitchen is no longer just a place of fire and steel. It is becoming a data-driven tech hub. The title of this deep dive is Line Cook to Code for a Specific Reason. Culinary professionals must become data fluent. To truly understand how to navigate this transition, we can actually draw some profound, unexpected lessons from the tech industry itself. Our sources included profiles from She Can Code, highlighting the journeys of self taught software developers. At first glance, a coder seems entirely disconnected from a chef. But the mental framework required to transition from a non-technical background into software engineering is the exact same mental framework a modern culinary professional needs to adopt to reach the C-suite today. Because a kitchen is essentially a biological computer, a recipe is an algorithm. Line cooks are compiling ingredients into outputs based on strict logical steps. Moving from understanding a recipe to understanding a software dashboard is A lateral shift in logic, not a leap across a canyon. Let's look at the specific stories from She Can Code to illustrate this mindset. Take Anina, for example. Anina's story is incredible. It is so grounding. She was literally mopping floors at a tanning salon. She didn't have a formal computer science degree or a pedigree from Stanford, but she recognized the shifting economy. So what did she do? She quit her tanning job, moved in with her aunt to lower her overhead, utilized free online coding resources. joined a rigorous boot camp, and networked relentlessly at local meetups until she landed a job as a software engineer at O'Reilly Media. It's the definition of grit applied to modern skill acquisition. Or consider Austin from our sources. He graduated with a chemical engineering degree, but realized his true passion was software. Right. While working exhausting full time shifts at an oil refinery, he spent nine grueling months diving headfirst into programming. He set strict daily goals, Built a foundational understanding of computer science principles and ground his way into an engineering role. We also have Victor, whose story is a perfect cautionary tale for this transition. When Victor first tried to build an Android app, his strategy was just to copy and paste code he found online. Because he just wanted the shortcut. Exactly. He wanted the end result instantly. But the result was a buggy, unstable application that crashed constantly. He realized that skipping the fundamentals doesn't work. He had to stop, go back. and focus intensely on the process of learning the underlying architecture, not just rushing to produce an output. That is the exact equivalent of a young chef trying to put molecular gastronomy foams on a menu before they even know how to properly roast a chicken. The foundation matters. It really does. And finally there's Anton, who pulled off his entire self taught developer journey without spending a single dime on formal education, relying purely on extreme resourcefulness, reading documentation, And engaging with online communities like Reddit. What is so vital about these tech transition stories is how directly they apply to the modern kitchen. If a talented Sioux chef wants to rise beyond the hotline and become a multi-unit culinary director or a C suite executive, having incredible knife skills is no longer enough. Being able to break down a whole pig is impressive, but it won't save a three percent margin. No, it won't. They must adopt that self-taught developer mindset. They need to become incredibly resourceful like Anton, they need to focus on the deep processes of data collection, like Victor, and they must become fluent in the language of business, EBITDA, supply chain logistics, and AI technology adoption. This evolution of the skill set is codified in what our sources call the Food Services 3 educational curriculum. The academic frameworks that train the next generation of chefs have completely changed. Modern programs expect advanced culinary leaders to Act as true holistic managers. It is no longer just about making a perfect glossy demiglaze. The modern culinary leader must possess advanced serve safe compliance knowledge, they must oversee and troubleshoot complex digital equipment, they have to design and conduct rigorous sanitation audits, and they must run a kitchen using managerial-level data-driven decision making. The curriculum explicitly dictates that these leaders have to balance their artistic creativity with hard cost efficiency. marketing strategy and evolving guest expectations. They are expected to research, develop, theoretically cost and execute menus while simultaneously training, scheduling, and delegating to their teams using algorithmic tools. The intellectual gap between an executive chef and an algorithmic systems manager is closing rapidly. They are basically becoming the same job. So for the operators and leaders listening right now who recognize they need to make this transition, How do they actually execute it without tearing their business apart? Our sources lay out a highly specific four-step implementation blueprint. Step one is the absolute foundation: establish ground truth telemetry. Right, because you can't fix what you can't measure. This means connecting all of your disparate digital systems. You must integrate your point of sale system, your kitchen display screens, your computer vision inventory scales. And your OCR invoice capture tools so they are all communicating and generating a single real-time operational feed. If your data is siloed and messy, applying AI will just help you make terrible decisions much faster. Exactly. Once the telemetry is clean, you move to step two. Delegate high variance low empathy tasks to the AI. I love that phrasing. Low empathy tasks. You do not use AI to automate the hospitality. You you don't use a robot to ask a guest about their anniversary. No, definitely not. You use AI to automate the soul crushing math. You hand off the dynamic par ordering, the weather forecasting, the delivery app throttling, and the inventory reconciliation to the autonomous agents. Which naturally leads to step three. Redefine the cook's role. With the math delegated to the machine, you must actively shift the human labor away from frantic calculations under stress. You move the human focus toward tactical, High quality culinary execution, guided by the dynamic display prompts the AI provides. You let humans do what humans do best: taste, touch, and care. And finally, step four, which is the most critical for actually getting the staff to buy in, tie variable incentives to system adherence. This is where you put your money where your mouth is. Exactly. You must align your crew's financial interests with the success of the technology. If the AI waste reduction scales and the automated ordering algorithms save the restaurant four thousand dollars a month in prime costs, the staff should share in that upside. You tie their performance bonuses to their waste reduction scores and their digital order completion accuracy. When the system makes the restaurant wealthier, the workers must become wealthier. It builds trust. Okay. I have to push back here. This four step plan sounds incredibly elegant in a theoretical textbook, but let's look at the reality on the ground. One of our sources explicitly warns about something called SaaS fatigue. Software is a service fatigue. yeah, it is a massive roadblock for independent operators right now. If an independent restaurateur is barely surviving on a three percent net margin, they are already drowning in monthly software subscriptions. They are paying a monthly fee for their POS, another fee for their loyalty app, another for their review management software, another for their scheduling tool, and they are bleeding out from the delivery app commissions. This death by a thousand cuts. Now we are telling them they need to buy a cutting edge agentic AI inventory platform that might cost an additional hundred and fifty to four hundred dollars a month per location? How does a midsized operator afford to stack yet another software fee without going bankrupt before the technology's return on investment ever kicks in? It is the ultimate catch twenty two of modernization. It's the pennywise pound foolish margin squeeze. You can't afford the tool that will save you money. So what's the workaround? The solution, as strongly recommended by the implementation experts at SurfStack and our sources, is to start small and scale smart. You do not buy a massive all-in-one enterprise AI suite on day one and try to rip out all your old systems. Because the implementation shock will just kill the business. Exactly. You start by digitizing the foundation you already have, usually the point of sale system. Ensure it is cloud-based and API ready. Then You aggressively identify your single biggest operational pain point. Where is the most cash bleeding from the business? Okay. So if your forensic account analysis shows that your biggest leak is food spoiling in the walk-in, you ignore the AI scheduling for now and you focus purely on installing the computer vision waste tracking. Yes. Or if your food costs are actually tight, but your labor budget is exploding due to overtime and turnover, you focus solely on AI demand forecasting and scheduling. So you solve one bleeding wound at a time. Precisely. And most importantly, when you do select a vendor, you demand integrated solutions. You stop buying six separate tools that refuse to talk to each other. You look for modern vendors that build agentic AI capabilities directly into the existing ecosystem. Aaron Because the ultimate goal is consolidation. Yes. If you can replace three clunky legacy subscriptions with one centralized agentic platform, you not only fund the new technology. But you drive a measurable permanent drop in your prime costs. You make the technology earn its keep before you expand it. It's a pragmatic, step by step evolution rather than a reckless revolution. Looking back at the massive landscape we've covered today, from the smart scales weighing the carrot peels to the algorithms dynamically setting the price of a burger based on the weather, to the restructuring of the dish pit into a technical apprenticeship, I want to leave you, the listener, with a final provocative thought to mull over. If agenc AI is standardizing the recipes, if it is calculating the exact portion costs to the penny, if it is predicting the daily foot traffic and autonomously ordering the raw ingredients from the suppliers before the chef even knows they need them, what does the executive chef of twenty thirty five actually look like? That is the big question. Will the ultimate marker of hospitality talent no longer be how fast you can perfectly dice an onion on a cutting board, but rather how How fluently you can communicate with and lead the algorithms that are truly running your kitchen. It is a profound paradigm-altering shift. The kitchen of the future does not need a human calculator screaming over a hot stove. It needs a deeply empathetic human leader, understands the code, understands the data, and uses those tools to elevate the human experience of dining. We look forward to having you join us for our next episode of From Kitchen to C Suite. Where we will continue to explore captivating topics designed to elevate your career and entrepreneurial drive. Don't miss a beat. Whether you're commuting or chilling on the couch, we've got you covered. On the go. Subscribe and listen via Riverside and from Kitchen2C Suite.com or our website hrbuni.com. You can also search for us on Apple Podcasts, Spotify, YouTube, Amazon Music Audible, iHeartRadio, or Deezer, or wherever you get your daily audio effects. Take HRB Network wherever you go. Download the new mobile web app, HRBUNI.com forward slash APP. On the big screen, level up your experience by watching the From Kitchen to Sea Suite podcast on the HRB Network on Roku, Google TV, or Fire TV. On social media, at HRBuni or at HRB Universal. Until our next broadcast, continue to innovate and strive for superior achievement.