McDonald's AI Pricing Push Meets a Firm Denial About Who Sets Menu Prices
McDonald's AI pricing system now guides recommendations across nearly 14,000 U.S. restaurants, according to a Reuters investigation published September 29. Yet McDonald's says artificial intelligence never sets prices, changes them, or determines what any individual customer pays.
That distinction defines the controversy. Reuters describes a machine-learning engine that analyzes millions of daily transactions and recommends an “optimal price” for each item at each restaurant. McDonald's describes the same technology as an advisory tool that leaves every final decision with human franchisees.
Both descriptions can be true. An algorithm can influence a decision without executing it. A recommendation can also carry more weight when the company tracks whether restaurant owners follow it.
The real question is therefore not whether software presses a digital button. It is how much practical control an AI recommendation gains inside a franchise system where corporate headquarters holds substantial leverage.
That question matters beyond hamburgers. Companies are adopting algorithms to turn demand, competition, and customer behavior into pricing advice. McDonald's now offers a prominent test of whether consumers will accept that process when affordability is already a sensitive issue.
What McDonald's AI Pricing Actually Does
The system reportedly recommends restaurant-level prices, while franchisees retain formal authority over the final menu.
Reuters reports that McDonald's has used some form of AI pricing technology since at least 2019. Its current engine continuously analyzes transaction data from the chain's extensive U.S. restaurant network.
Machine learning means software identifies patterns in historical data and uses those patterns to make predictions or recommendations. In this case, the output is a suggested price for each menu item at a particular restaurant.
The inputs reportedly include local demand, operating conditions, competitor menu prices, and estimates of customer price sensitivity. Screenshots reviewed by Reuters referenced “customer willingness to pay” within a restaurant's area.
That phrase prompted understandable concern. It suggests a system designed to identify how much more a local market might tolerate before sales begin falling.
However, the reported tool operates at the restaurant level, not the individual customer level. There is no verified evidence that McDonald's presents different base menu prices to two people ordering simultaneously at one location.
McDonald's reinforced that distinction in an October 1 pricing clarification. The company said its system does not use personal information to determine what an individual customer should pay.
McDonald's also denied using dynamic pricing. Dynamic pricing changes prices in response to conditions such as demand, timing, inventory, or capacity. Airlines and ride-hailing platforms commonly use versions of this approach.
The company says its tool does not change prices in real time or charge more during specific hours. Instead, it produces restaurant-specific recommendations based on geographic markets and other local conditions.
That makes “AI-assisted local pricing” more accurate than “personalized pricing” or “surge pricing.” It also makes the system less dramatic than some headlines suggest.
Still, the recommendations appear more sophisticated than an occasional spreadsheet. The Reuters investigation says McDonald's sends pricing guidance to franchisees at least three times annually.
Reuters also reports that Tiger Analytics runs the platform and works with McDonald's on its parameters. Former Tiger employees said those parameters reflected corporate goals, including attracting customers or improving profits.
The reported rules were not designed only to raise prices. Some prevented increases on selected products during particular seasons. Others focused proposed increases on items whose prices had remained unchanged for longer periods.
More recently, franchisees reportedly received recommendations for conservative pricing and some reductions. That change reflects McDonald's effort to rebuild its reputation for affordability after several years of rising restaurant costs.
The system therefore does more than search for the highest tolerable number. It reportedly balances sales volume, customer traffic, operating costs, competitive positioning, and corporate priorities.
Yet that broader purpose does not eliminate the controversy. It shifts attention toward the relationship between corporate recommendations and franchisee decisions.
The Tool Versus Mandate Dispute
McDonald's says the engine provides optional advice, but several franchisees told Reuters that rejecting its recommendations brings pressure.
McDonald's operates through a heavily franchised model. Approximately 95% of its U.S. restaurants are owned and operated by franchisees, according to value-menu reporting.
Those owners formally set their restaurant prices. They also carry local expenses, including wages, rent, utilities, insurance, food, and packaging.
Corporate headquarters has different economic incentives. McDonald's receives revenue connected to restaurant sales, while each franchisee must protect the profitability of an individual location.
A lower price can attract more customers and raise total sales. However, it can also weaken a restaurant owner's margin when local costs remain high.
That tension predates artificial intelligence. The algorithm gives headquarters a more consistent, data-rich way to recommend how franchisees should resolve it.
McDonald's says a recommendation remains exactly that. Its public response states that restaurant owners can reject the tool's output and independently determine their final prices.
Reuters found evidence of a more complicated operational reality. Five franchisees described pressure to use the recommendations, although another former owner did not feel compelled to accept them.
One document reviewed by Reuters reportedly tracked deviations from the system's guidance. Since January, franchisees have also been required to engage constructively with approved pricing consultants and tools under updated business standards.
Corporate conversations can matter because McDonald's evaluates franchisees on issues affecting their future within the system. Those decisions can include restaurant renewals and opportunities to open additional locations.
CEO Chris Kempczinski acknowledged in August that pricing noncompliance can enter certain franchisee discussions. That statement does not prove that McDonald's dictates prices, but it shows that disagreement can receive corporate attention.
This creates a gap between legal authority and practical influence. A franchisee can possess the formal right to reject a recommendation while facing business incentives to follow it.
That is the core dispute behind McDonald's AI menu prices. Focusing only on who enters the final number misses how organizational power shapes the decision.
The same issue appears throughout enterprise AI. Companies often describe automated systems as decision support, with a human remaining in control.
That safeguard only works when the human has meaningful freedom, sufficient information, and a realistic ability to disagree. A nominal approval step does not automatically establish independent judgment.
McDonald's rejects the idea that its standards amount to control over final prices. The company says its assessments consider the broader customer experience, not just whether a restaurant follows pricing guidance.
The available evidence does not establish that every franchisee receives identical pressure. It also does not show that restaurants automatically implement every recommendation.
It does show that the pricing engine operates inside a relationship with unequal bargaining power. That makes adoption rates and deviation records important indicators, not administrative details.
Why McDonald's Is Expanding AI Price Recommendations Now
The company needs prices low enough to restore customer traffic, but high enough to sustain franchisee economics after years of cost inflation.
McDonald's entered this debate with an existing affordability problem. The company said its average U.S. menu prices increased about 40% between 2019 and 2024.
McDonald's attributed those increases to comparable growth in major restaurant inputs. Its pricing fact sheet said crew wages rose about 40%, while food and packaging costs increased about 35%.
The company argued that its overall increases remained consistent with higher operating expenses. Consumers still experienced the result as a more expensive visit.
Individual restaurant prices also vary. Reuters found a 21% difference between Big Mac prices at two company-operated Fresno locations only two miles apart.
That comparison proves geographic price variation, but not its cause. Reuters explicitly said it could not determine whether the AI engine produced that difference.
Local pricing existed before machine learning. Restaurants near airports, highways, tourist areas, and costly urban centers have long charged differently from suburban locations.
McDonald's says nearby restaurants can belong to distinct markets with different expenses and demand. An algorithm can measure those differences more precisely, but it did not invent them.
The complication is that greater precision can widen variation. Three franchisees told Reuters that the engine expanded existing price differences between restaurants, including locations in nearby neighborhoods.
For consumers, the source of a difference might matter less than its perceived fairness. A customer comparing two app menus sees the same product and brand carrying different prices.
If the higher price appears connected to unavoidable costs, customers might accept it. If it appears designed around how much a neighborhood can tolerate, the reaction changes.
McDonald's also faces pressure from lower-income customers who have reduced visits across the quick-service sector. Reuters cited Placer.ai estimates showing declining year-over-year U.S. traffic at McDonald's during every complete month since March.
The company has responded by emphasizing affordability. Its updated national value strategy includes a simpler selection of lower-priced items and meal offers.
That strategy requires cooperation from thousands of independent operators. About one-third of franchisees did not follow certain recommended value-menu pricing, Kempczinski said during an August earnings call.
He said those restaurants produced softer business results, although the company did not disclose enough evidence to independently test the claim. McDonald's can use its engine to argue that lower prices generate enough additional traffic to offset reduced unit margins.
Franchisees may reach a different conclusion after examining their own labor, rent, and food expenses. Their objective is not maximizing systemwide sales at any cost.
This explains why McDonald's AI pricing is becoming more important now. Headquarters needs a repeatable method for identifying where lower prices can restore demand without sacrificing too much revenue.
The technology promises more specific guidance than a nationwide directive. It can recommend reductions in some markets, hold prices elsewhere, and identify products that appear less sensitive to increases.
That precision may help McDonald's support a consistent value message while acknowledging local economics. It can also give headquarters a quantitative argument when a franchisee prefers a higher price.
The technology is therefore not merely a pricing calculator. It is becoming part of how McDonald's negotiates affordability across its franchise network.
McDonald's AI Pricing Is Not Dynamic Pricing, but Trust Still Matters
McDonald's rejects the dynamic-pricing label, yet consumers can still question recommendations built around local willingness to pay.
The difference between restaurant-level recommendations and real-time personalized pricing is substantial. McDonald's says everyone ordering from one location sees the same applicable base menu price.
Its system does not reportedly raise lunch prices because a line forms at noon. It does not lower them automatically during an empty afternoon.
McDonald's also says the engine does not inspect one person's purchasing history and choose a higher menu price for that shopper. Nothing in the verified reporting establishes such individualized pricing.
These limits separate the reported practice from the most sensitive forms of algorithmic pricing. They should not be blurred simply because all three approaches can involve artificial intelligence.
However, the absence of real-time or individualized pricing does not settle the fairness question. A neighborhood-level estimate of willingness to pay still divides markets using inferred behavior.
That estimate can reflect income, competition, commuting patterns, transportation access, or the availability of alternatives. Consumers rarely know which variables shaped the recommendation.
The tool reportedly uses publicly available competitor prices from nearby restaurant menus. Wendy's and Burger King told Reuters they do not use AI in their pricing decisions.
Competitor data can help a restaurant avoid moving far outside its local market. It can also create concerns if many businesses use algorithms that respond to each other's prices.
The legal issue is not that software processes public information. Regulators focus on whether pricing systems facilitate coordination, discrimination, deception, or other prohibited conduct.
Reuters reviewed portal terms warning users that franchisees might be competitors. The terms reportedly instruct them to follow competition laws and consult their own lawyers when necessary.
William Kovacic, a former Federal Trade Commission commissioner, told Reuters that the warning acknowledges a potential issue. Other legal experts assessed the risk as relatively low because courts allow franchise brands considerable pricing influence.
No regulator cited in the reporting concluded that McDonald's violated antitrust law. The FTC and Justice Department did not comment for the Reuters article.
The broader regulatory environment is becoming less forgiving. The FTC has warned companies that undisclosed use of personal data in individualized pricing can create consumer-protection concerns.
McDonald's says it does not engage in that practice. The warning still explains why precise terminology and public transparency have become essential.
Wendy's learned the reputational risk in 2024. Comments about digital menu boards and pricing experiments triggered widespread claims that the chain planned surge pricing.
Wendy's denied that interpretation and said it had not implemented such a system. The episode showed how quickly consumers associate algorithmic pricing with paying more during busy periods.
Instacart faced related criticism after testing tools that displayed different grocery prices to different shoppers. It later ended that limited test and said personal information would not determine item prices.
McDonald's has responded with categorical language. AI does not set menu prices, the company says, and people remain responsible for final decisions.
That answer addresses automation but only partly addresses influence. Consumers may also want to know how recommendations are developed, which variables are excluded, and how often franchisees reject them.
Transparency should distinguish base prices from personalized discounts. A loyalty offer can vary by customer without changing the posted price everyone sees.
Businesses must explain that distinction clearly. Otherwise, personalized promotions, geographic pricing, and individual price discrimination can merge into one alarming public narrative.
McDonald's should also explain how it tests recommendations for unintended geographic disparities. A system can avoid personal data while still producing outcomes correlated with sensitive demographic characteristics.
No public evidence currently shows that its engine systematically disadvantages protected groups. It would be irresponsible to claim otherwise without location-level analysis.
The verification gap still matters. McDonald's has described what the tool does not do, but it has released limited detail about its models, governance, inputs, or auditing process.
That leaves consumers choosing between a corporate assurance and an investigation based on internal screenshots, documents, and anonymous franchisee accounts. More disclosure would make the distinction easier to evaluate.
The Bigger Contest Is Corporate Optimization Versus Local Judgment
The central conflict is not McDonald's against another restaurant chain. It is algorithmic consistency against franchisee control and consumer trust.
McDonald's can view the same data across thousands of restaurants. That scale lets its system detect patterns that one operator cannot see from a handful of locations.
An engine can estimate how a price change affected transactions in comparable markets. It can separate seasonal movement from a restaurant-specific problem more effectively than intuition alone.
The company can also coordinate national value messaging. Advertising becomes difficult when promoted prices differ widely or too many franchisees decline the offer.
Franchisees contribute information that a centralized model may not capture quickly. A local operator understands staffing shortages, construction disruptions, neighborhood events, and competitive changes that historical transactions may miss.
This does not create a simple contest between humans and machines. The useful model combines broad data with accountable local judgment.
Trouble begins when either side treats its information as complete. Corporate teams can overestimate the model's precision, while restaurant owners can mistake recent experience for a durable pattern.
The AI system also changes the burden of proof. Before such tools, headquarters had to persuade an owner that a different price would perform better.
Now the recommendation arrives with the authority of millions of transactions and a calculated optimum. The owner who disagrees must explain why one location is an exception.
That asymmetry can improve discipline. It can also create automation bias, which occurs when people defer to software because its output appears more objective than it is.
An “optimal price” is never neutral. Optimization requires a target, constraints, time horizon, and assumptions about acceptable tradeoffs.
A model prioritizing transaction growth can recommend something different from one prioritizing restaurant profit. A system protecting market share can tolerate outcomes that burden some franchisees.
Reuters reports that McDonald's has supplied corporate goals and rules to the platform operator. That human direction is important because the algorithm does not independently decide what success means.
McDonald's says affordability and customer value remain central. Franchisees may support that goal while disputing how its costs should be distributed across the system.
Consumers introduce a third objective. They want predictable, understandable prices and assurance that the company is not exploiting limited local alternatives.
A technically effective model can still fail if customers see its recommendations as unfair. Price optimization depends on behavior, and public distrust changes behavior.
McDonald's previous AI experience offers a useful warning. The company ended an IBM automated drive-through ordering test in 2024 after deployments at selected restaurants.
The drive-through experiment attracted attention when customers shared ordering errors online. McDonald's continued exploring AI, but the episode showed that scale does not guarantee a technology is ready.
Pricing creates an even more sensitive challenge. A mistaken order causes frustration, while a seemingly unfair price can damage the entire brand's value promise.
The engine's output must therefore survive more than statistical testing. It must remain defensible to operators, customers, regulators, and the public.
For enterprise buyers, McDonald's offers a general lesson. Human oversight cannot be measured only by whether someone approves the final output.
Organizations should examine rejection rates, escalation paths, incentives, and consequences. If users almost never reject an algorithm, leaders should determine whether the model is consistently right or disagreement feels unsafe.
They should also document the objective being optimized. Terms such as “optimal,” “recommended,” and “customer value” can conceal competing definitions.
The best governance asks who benefits when the system succeeds and who absorbs the cost when it fails. At McDonald's, those answers differ across corporate headquarters, franchisees, and customers.
Three Signals Will Show Where the Strategy Goes Next
Adoption data, customer traffic, and regulatory responses will reveal whether McDonald's can turn pricing intelligence into trusted business guidance.
The first signal is franchisee behavior. McDonald's does not need to publish proprietary model code, but it can provide aggregate acceptance and rejection rates.
Those rates would clarify whether the platform functions as optional advice. A high acceptance rate proves little by itself, but unexplained increases following compliance requirements would deserve scrutiny.
McDonald's could also disclose how often restaurant owners modify a recommendation. Meaningful adjustment would support its claim that people retain control.
The second signal is customer traffic alongside value perception. McDonald's is using more conservative recommendations while promoting affordability, so the strategy should produce measurable behavioral improvement.
Higher transactions without stronger value perceptions would suggest promotions are doing temporary work. Better perceptions without stronger traffic would suggest affordability remains constrained by other factors.
Results must also be compared with franchisee economics. A traffic gain is not durable if participating restaurants cannot cover their operating expenses.
The third signal is regulatory guidance on algorithmic and personalized pricing. McDonald's says its restaurant-level approach does not personalize base prices, yet regulators are closely examining adjacent practices.
Any new disclosure expectations could push companies to explain what data enters their pricing systems. Rules may also clarify when franchisees count as competitors for algorithmic-pricing analysis.
These signals will determine whether McDonald's AI pricing becomes a model for responsible decision support or a lasting source of distrust.
For now, the strongest conclusion is narrower than the original headline. AI reportedly helps calculate and recommend local menu prices, but McDonald's says humans decide what appears on each menu.
The unresolved issue lies between those statements. A human decision is meaningful only when the person can understand the recommendation, challenge it, and reject it without improper consequences.
Customers should watch what McDonald's discloses, not merely what it denies. Franchisees should demand clear objectives, audit trails, and credible freedom to exercise local judgment.
Other companies deploying pricing algorithms should ask the same questions before controversy arrives. Does the system advise, influence, or effectively control the decision?
McDonald's now has an opportunity to answer that question with evidence. Until then, its AI pricing engine will remain both a business tool and a test of how much algorithmic influence consumers will accept.
