Sooner or later, three different people will ask you some version of the same question.
A food safety investigator will want every lot of leafy greens that moved through your kitchens over the past two weeks: where each case came from, what you made with it, and where the finished food went. A client’s auditor will want proof that every meal you billed last quarter was made to spec and actually served. An analyst at a firm that wants to invest in your company will want food and labor cost per meal, by site, for the last three years.
The answers exist. They are scattered across point-of-sale systems that were never introduced to one another, production sheets on clipboards, a spreadsheet a regional manager rebuilds every month, and the memory of forty kitchen managers, one of whom retired in June.
The food is the easy part.
A Different Business Than It Appears
From the outside, a company that runs dozens of commercial kitchens can look like a restaurant group without the storefronts. It is a different business. If you run contract dining, catering, a commissary, or a central production kitchen, the person eating is usually not your customer. Your customer is a hospital, a senior living community, a university or school, a corporate campus, a stadium, an airline, or a restaurant brand that has handed its prep to you. Your price is set by a contract rather than a menu board. Your kitchen often sits in someone else’s building, runs on someone else’s equipment, and answers to someone else’s rules, with a crew you hired locally.
Talent still matters more than anything else in a kitchen. It just does not scale the way a system does, and it does not show up in a data room.
That makes every kitchen a branch. Each one has its own client, its own contract terms, its own equipment, its own staff, and increasingly its own regulatory footprint. The company is a multi-branch business, and multi-branch businesses in banking, retail, and health care learned long ago that branches run on shared systems or they stop scaling. Kitchen operators could postpone that lesson longer than most, because talent covered the gap. A strong chef and a good general manager could carry a site. A strong regional director could carry ten. Talent still matters more than anything else in a kitchen. It just does not scale the way a system does, and it does not show up in a data room.
This is why so much of the technology advice aimed at food service misses these operators. The digital front door, dynamic menu pricing, and the contest with third-party delivery apps matter enormously to restaurants. They barely touch a company whose customer is a procurement committee.
Why the Pressure Is Rising Now
Three forces are converging on the back of the house.
The first is traceability. The FDA’s Food Traceability Rule, which implements Section 204 of the Food Safety Modernization Act, covers the foods on the agency’s Food Traceability List, including leafy greens, fresh-cut produce, soft cheeses, shell eggs, and seafood. Companies that handle them must keep lot-level records at key steps and, when the agency asks during an outbreak or recall, produce those records as a sortable electronic spreadsheet within 24 hours. It reaches restaurants and institutional food service operators, not only processors and distributors, with some exemptions for the smallest operators. Enforcement now waits until July 2028, which sounds like a long runway until you consider what the rule assumes: that your receiving, production, and shipping records live somewhere a query can reach them.
The second is the ingredient list. States have begun writing their own rules about what goes into food, and the rules do not line up. California’s ban on a short list of additives takes effect at the start of 2027. Louisiana has enacted disclosure rules, including one that will require food service establishments to tell customers when they cook with certain seed oils. Other states’ laws are tied up in court, and new bills arrive every session. Meanwhile, the federal Dietary Guidelines released in January 2026 are the first to urge Americans to limit highly processed foods, and those guidelines steer the federal meal programs whose standards many institutional menus follow. Cooking from whole ingredients answers that shift, and it multiplies the lots, suppliers, and recipe steps you have to track. Which recipe, at which site, uses which ingredient is no longer a culinary detail. It is a compliance question.
The third is capital. The largest contract caterers have spent years buying regional operators, and private equity has been assembling platforms of its own. Buyers pay for what they can verify. A company whose numbers live in forty kitchens is harder to value than one whose numbers live in one place, and the difference shows up in the price.
Each of these forces asks the same thing of your kitchens: that every one of them can pass four tests. Call them the four Cs: counted, compliant, consistent, and costed. The order is deliberate, because each test leans on the ones before it.
Test One · Counted
In most of this business, the count is the invoice. You bill per meal, per head, per patient day, per event, or per claim, and every one of those counts begins where food meets a person. When the point-of-sale record at a site does not reconcile with the production record, and neither reconciles with what the client was billed, the errors run in both directions: meals served and never billed, which is lost revenue, and meals billed that a client’s auditor will later dispute, which is lost trust.
The traceability rule adds a second kind of count. From 2028, the lot record is the trace. The discipline that proves what you served is the same discipline that proves what you received and where it went.
This is the unglamorous first step, and it comes before anything clever: one source of truth for production, service, and billing across every site. In our Stabilize, Optimize, Monetize sequence, it is the Stabilize phase, and skipping it is the fastest way to stall an AI initiative in a multi-site operation. You cannot automate a reconciliation you have never managed to do by hand.
Test Two · Compliant
Compliance in a commercial kitchen used to mean a clean inspection and a binder of food safety plans. It now means being able to show, on demand, that every recipe at every site meets a stack of requirements that varies by jurisdiction and by client: allergen controls, traceability records, nutrition standards where they apply, ingredient restrictions that differ from state to state, and whatever the client added to its spec at the last renewal.
No dietitian can hold all of that in memory across a network of kitchens, and no one should have to. Once the counts live in one place, the rules can live there too. When the recipe, the ingredient database, and the supplier catalog are connected, a change in one state’s law becomes a list of affected recipes by site in minutes rather than a month of phone calls. This is where AI proves its value early: checking every recipe against the current rule set, watching for regulatory changes as they happen, and assembling the traceability spreadsheet from receiving records you already keep. Rules built into the system work the way guardrails work on a freeway. They are what let a kitchen move fast without looking over its shoulder.
Test Three · Consistent
The promise you sell a client is that the dish on the menu is the dish on the plate, at every site and on every shift. Forty cooks will produce forty versions of the same recipe unless something holds them to one. For years that something was a chef walking the line and a regional director who could get to each kitchen twice a month.
Vision AI adds a set of eyes that does not tire. A photo at the pass, or a camera over the tray line, can be checked against the recipe specification for components, portion, and presentation, with the exceptions routed to the person who can fix them. Tray-line cameras built for hospital and skilled nursing kitchens already check each tray against a patient’s diet order and keep an image of every one. The same idea serves a campus kitchen, a catering line, or a commissary shipping to twenty restaurants, and the photo record doubles as evidence when a client asks what was served.
Two cautions keep this from going wrong. The first is oversight that is real rather than ceremonial. A system that flags too much trains people to click through the flags, and a reviewer who approves everything is a rubber stamp, not a safeguard. The second is privacy. Photograph the plate, not the people. A camera that captures faces in a dining room turns a quality tool into a privacy exposure, and one that identifies them can become a legal problem in states with biometric privacy laws. The disciplined path is to let a few curious managers prove the idea as Citizen AI, then promote what works into Production AI with real engineering discipline behind it.
Test Four · Costed
A commercial kitchen’s margin lives in precision. Overproduce, and the margin goes into the waste bin. Underproduce, and you have a service failure in front of a client who can put your contract out to bid. Schedule labor to last year’s pattern, and you overpay on the slow days while burning out your crew on the busy ones.
The operators pulling ahead forecast demand the way airlines forecast seats. They use the calendar, the menu, the weather, the client’s events, and the history of every site to set production and staffing, then measure waste instead of guessing at it. They know food and labor cost per meal by site, by week. That last number matters well beyond the kitchen. It is one of the first figures a buyer’s diligence team asks for, and one that many operators can produce only by quarter, by region, after a month of spreadsheets.
This is the Optimize phase, and it depends on everything before it: the counts from the first test, the recipe data from the second, and the portion control from the third.
What the Buyer Checks First
The AI dividend is the gap between what work used to cost and what it costs now. Our 2026 AI Dividend Map asks what shape that dividend is taking in each industry, which decides who gets to keep it. Most of food service, restaurants above all, sits under a Rising Bar, one of the four shapes of the AI dividend: whatever AI saves an operator, its customers soon expect as standard, so the gains have to be reinvested to stay ahead.
Kitchens that cook under contract behave differently. In the map’s terms, the client can spec the work and can shop it, but the market will not sole-source it, because the food has to be made near where it is eaten. That combination is what the map calls a Rollup. When a market will not consolidate on its own, capital consolidates it by acquisition.
In a Rollup, the AI dividend shows up in your multiple before it shows up in your P&L. The platform that can bring an acquired operator’s kitchens onto one recipe system, one point-of-sale standard, and one billing and traceability record in weeks rather than a year is the platform that can keep buying. The one that cannot is the one being bought, and priced accordingly. This is the Monetize phase of our sequence, where technology starts to show up in revenue and enterprise value rather than in cost alone.
A buyer’s diligence team will run the four tests whether you do or not. So will the next client that puts your contract out to bid. It is better to run them yourself, and first.
Where to Start
The order matters more than the tools. Unify the counts first, then put the rules into the system, then add the vision and forecasting capabilities that depend on both. An independent assessment of your technology platform, led by someone who has run technology in multi-unit food operations, will show you in a few weeks which of the four tests you are closest to passing and which one is quietly costing you the most.
From there, the work is leadership rather than software. Some multi-site operators have a technology leader who needs more strategic depth or bandwidth. Many have an IT manager who keeps the systems running but was never asked to carry a roadmap across forty sites. Either way, closing that gap is what our Contract CIO+® and CIO IQ® services are built for, with sector-matched technology leaders who have run operations like yours and who align technology to how you actually make money. More than half of the measurable impact we have driven for clients has been top-line growth rather than cost savings, and in this business the top line is won one contract at a time. See how we work with food service and commercial kitchen operators.
Every kitchen is a branch. The companies that run them that way will be the ones doing the buying.
Every kitchen is a branch. The companies that run them that way will be the ones doing the buying.
Every kitchen is a branch. The companies that run them that way will be the ones doing the buying. If you want a clear picture of which of the four tests your kitchens are closest to passing, see how we work with food service and commercial kitchen operators.
