AI is entering Miami’s restaurant scene. As the city’s restaurateurs struggle to make sense of the overwhelming data and cost leaks that make this one of the most difficult industries to succeed in, some are looking to an AI general manager for solutions.
Restaurant owners have never had more data at their fingertips. Between point-of-sale systems, payroll platforms, inventory software, labor scheduling tools, delivery apps, and accounting systems, operators are flooded with information about every aspect of their business.
Yet many still struggle to answer a surprisingly simple question each morning: Did we actually make money last night?
According to Saleem Khatri, CEO of Lavu, which created Marty, an AI general manager for restaurants, that’s because most restaurant technology was never designed to help operators make decisions. It was built to record transactions or report on what already happened.
“Marty isn’t another dashboard,” says Khatri. “It’s the operational brain that sits on top of all of a restaurant’s systems and tells owners exactly where money is leaking, who broke the rules, and what they need to do before the next shift starts.”
As artificial intelligence continues to reshape industries ranging from finance to healthcare, a growing number of restaurant operators are beginning to use AI not as a marketing tool or chatbot, but as a virtual general manager.
The Hidden Cost Of Operational Blind Spots
The restaurant industry operates on notoriously thin margins. Small inefficiencies in labor, food costs, inventory controls, or employee behavior can quietly erode profitability for months—or even years—before anyone notices.
Marty was built specifically to uncover those hidden losses.
The platform analyzes more than 50,000 operational signals per restaurant across POS systems, labor scheduling software, payroll platforms, inventory management tools, delivery services, and even spreadsheets. It then verifies the underlying financial data and translates it into recommendations that operators can act on immediately.
Recent analysis conducted across a restaurant group revealed approximately $2.9 million in hidden operational exposure. That included $1.24 million tied to labor inefficiencies such as overtime creep and scheduling gaps, $980,000 in revenue exposure from discounting and guest retention issues, and another $680,000 in compliance and control risks ranging from suspicious voids to improper employee access practices.
“This wasn’t money disappearing from failing businesses,” Khatri says. “These were restaurants that owners considered well run. The problem is that most operators simply don’t have visibility into the slow bleed that’s happening every day.”
What Restaurant Leakage Actually Looks Like
The concept of operational exposure may sound abstract, but the real-world examples are striking.
In one high-end steakhouse group in the Southeast, Marty analyzed more than 67,000 guest transactions and $8.2 million in sales. The system identified an estimated $95,000 to $175,000 in annual recovery opportunities.
Among the findings were more than $26,000 in annual overtime costs at a single location, a sharp increase in discretionary comping practices, and a 15 percent decline in late-night revenue despite guest traffic increasing.
In another instance, an underperforming month was traced to a single event coordinator who had quietly reduced private dining bookings following a dispute with management. The decline appeared in the numbers long before leadership suspected anything was wrong.
For a two-unit fast casual operator in Southern California, Marty identified more than $10,000 per month in recoverable value. The system surfaced dramatic labor fluctuations, unusual void activity from an individual employee, recurring payroll waste from habitual late clock-outs, declining performance among key menu items, and marketing spend that continued despite collapsing engagement metrics.
Khatri describes these discoveries as “death by a thousand cuts.”
“Most owners aren’t losing money because of one catastrophic problem,” he says. “They’re losing it through dozens of small operational issues that compound over time.”
Some findings are even more dramatic.
At an international hospitality group in the Middle East, Marty helped uncover what management later determined to be years of systematic theft involving a bookkeeper, food inventory losses involving kitchen staff, and revenue reporting inaccuracies that distorted the true profitability of food operations.
Once corrected, management estimated recurring losses of roughly $20,000 per month had been eliminated.
Beyond Dashboards
The rise of restaurant analytics is hardly new. Operators have access to more dashboards than ever before.
Khatri argues that dashboards are precisely the problem.
“The industry has become very good at showing owners charts,” he says. “What owners actually need is someone to tell them what the chart means and what action they should take.”
Rather than requiring operators to interpret dozens of reports, Marty delivers two primary outputs.
The first is real-time operational alerts throughout the day. A restaurant manager might receive a notification that sales are running significantly below historical norms and labor should be reduced immediately. Another alert could flag abnormal void activity from an employee or identify that a delivery platform has stopped receiving orders during peak hours.
The second is what the company calls the Morning Receipt—a concise daily operational summary delivered before opening. Instead of forcing owners to sift through reports, the system highlights the most important issues, quantifies potential savings, and recommends specific actions.
The goal is not to provide more information. It is to provide clarity.
“We don’t want restaurant owners spending hours analyzing data,” Khatri says. “We want them spending minutes understanding what matters.”
The Emergence Of The AI General Manager
Marty’s broader vision reflects a growing shift in how artificial intelligence is being deployed across businesses.
Rather than replacing workers, many of the most successful AI applications are functioning as force multipliers for management teams.
In restaurants, where operators often oversee multiple locations while juggling staffing, customer experience, food quality, and profitability, the need for additional oversight has become increasingly acute.
Marty integrates with major restaurant technology platforms including Toast, Aloha, SpotOn, PAR Brink, Restaurant365, MarginEdge, 7shifts, and others. Instead of requiring restaurants to replace existing systems, the platform acts as an intelligence layer that connects them.
That approach appears to be resonating. According to the company, Marty is experiencing double-digit weekly growth in usage, while customer retention remains strong as operators expand deployments across additional locations.
“We’re not trying to replace anyone’s technology stack,” Khatri says. “We’re trying to become the intelligence layer that makes all of those systems more valuable.”
From Personal Experience To Industry Solution
For Khatri, the motivation behind Marty is deeply personal.
He recalls watching his immigrant father build a successful business over many years, only to discover that a trusted accountant had allegedly stolen millions of dollars over more than a decade without detection.
Neither the bank nor external auditors identified the issue. By the time it was uncovered, the damage had already been done.
“My dad built a $20 million business and still didn’t have the visibility to see what was happening,” Khatri says. “A lot of restaurant owners are facing the same challenge today. They don’t know where the problem is until it’s already become expensive.”
That experience shaped the philosophy behind Marty.
Instead of building another reporting platform, Khatri wanted to create the analyst his father never had—one capable of continuously monitoring operations, identifying anomalies, and translating complex information into actionable decisions.
“Marty is essentially the analyst I wish my dad had,” he says. “The difference is that now we’re putting that capability in the hands of restaurant operators running two locations, five locations, or twenty locations.”
As AI adoption accelerates across the hospitality sector, the technology may ultimately be judged not by how much data it can process, but by how much uncertainty it can remove.
For restaurant owners navigating thin margins and constant operational complexity, that peace of mind may prove just as valuable as the savings themselves.
Written in partnership with Tom White
