McDonald’s Faces Questions Over AI Pricing
McDonald’s is facing scrutiny over its reported use of AI to recommend menu prices, raising questions about how algorithms influence what customers pay even as the company denies using personalised prices or changing charges throughout the day.
What’s Been Reported?
There have been recent allegations of McDonald’s using a pricing system that analyses transactions and local market information to recommend prices at individual restaurants. The reporting has focused on the US and some international markets, rather than establishing that the same arrangements apply throughout the UK.
The system reportedly considers how customers in an area respond to prices and includes publicly available competitor pricing. It’s also been reported that some franchisees, the independent operators running restaurants under the McDonald’s brand, say they feel pressure to follow its recommendations, although the company maintains that pricing decisions remain theirs.
What Does McDonald’s Say?
In a statement published on 1 October, McDonald’s rejected claims that (AI) software automatically determines what customers are charged, saying, “The tool provides information. People make the final pricing decisions.”
Its explanation describes recommendations based on conditions affecting particular restaurants, including geography, local costs, competition and demand. The company says operators decide whether to use those suggestions and are not required to accept them.
McDonald’s says restaurant operators make the final decision, but some franchisees reportedly feel under pressure to follow the suggested prices, raising questions about how much freedom they have to reject recommendations they believe are wrong for their customers or business.
Is This Really Dynamic Pricing?
Describing McDonald’s approach as ‘dynamic pricing’ could suggest that the price of a meal rises when a restaurant gets busy, much as fares can increase on ride-hailing apps during periods of high demand. That’s actually known as surge pricing, while personalised pricing is different, i.e., using information about an individual customer to decide what they are charged.
However, McDonald’s says it does neither, explaining that its tool “does not change prices in real-time or at different hours during the day” and does not recommend prices for individual customers. Instead, it says the recommendations reflect conditions at each restaurant, where rent, wages and competition can vary even between nearby branches.
A price difference between two restaurants therefore doesn’t establish that either is using surge or personalised pricing. The question raised by the reporting is how much influence AI recommendations have on those local prices and whether operators feel able to choose a different price when they believe it would better suit their customers.
Why Use AI To Recommend Prices?
For a large restaurant network, pricing involves many decisions across products, locations and changing costs. Analysing sales patterns can help operators estimate whether an increase would reduce purchases or whether a discount might attract enough extra customers to be worthwhile.
Used well, that could help a restaurant preserve affordable options while covering higher operating costs. There’s no reason a recommendation must always involve an increase, because lowering a price can sometimes generate more business.
However, the result depends on what the system is asked to achieve. Increasing revenue, improving profit and encouraging repeat visits are related goals, but a price that helps one may undermine another.
What Can Sales Figures Miss?
A completed purchase shows that someone accepted a price on that occasion, but it doesn’t really establish that they considered it good value or that they intend to return. In fact, someone buying food during a rushed journey may pay more than they would when choosing where to eat near home.
That creates a potential weakness if recommendations focus too heavily on existing transactions. Customers who have stopped visiting are no longer generating purchases, even though understanding why they left could be essential to maintaining the business.
Operators therefore really need to assess prices against repeat visits, customer feedback and profit after costs, rather than treating an immediate increase in takings as proof that a decision worked.
Who Has The Final Say?
The reports (allegations) of pressure on franchisees raise questions about how freely they can set prices, although McDonald’s says it assesses the overall customer experience rather than price alone and states, “Franchisees are not required to accept a pricing recommendation.”
For that freedom to actually work in practice, operators need to understand why a price has been suggested and be able to reject it based on their knowledge of local customers and costs. The same applies to any business using automated advice, where human approval only provides useful oversight if the person involved can question the recommendation and act on their own judgement.
What Does This Mean For Your Business?
Businesses considering pricing software should first define the result they want, whether that is sustainable profit, more repeat customers or better use of spare capacity. A limited trial should measure those outcomes over time, including whether an apparent improvement simply reflects customers buying sooner or switching to cheaper products.
Suppliers should explain which information feeds their recommendations and how uncertainty is handled. Managers also need clear authority to reject or reverse changes, with records showing why decisions were made. Where software uses information from other businesses, its sources and permitted uses deserve scrutiny before recommendations become routine.
The key point here is really that customers judge the price they encounter, not the sophistication of the system behind it. AI can support better decisions, but businesses should remain able to explain their approach and recognise when a short-term gain risks weakening the trust that brings customers back.
