Mercadona invests heavily in AI. And it has an advantage built over 40 years
The general overview. Mercadona has a turnover of 38.8 billion euros annually without being the cheapest supermarket in Spain. The slogan ‘Always Low Prices’ was left behind a long time ago because its focus became different. Their secret today is not in the price, but in constantly anticipating what their customers need. Now it wants to scale that capability with AI. Why is it important. Sergio Pajares, the company’s Director of Technology, sums it up bluntly in statements to The Spanish: The company does not seek to have AI for the sake of having it, but rather to develop “the best AI to sell lettuce.” A way to make it clear that they do not want a simple technological showcase, but rather tools that solve tangible problems of daily business. The context. Mercadona has been optimizing its supply chain for forty years and accumulating knowledge about what works in its stores and logistics blocks. That advantage built year after year is what now powers its AI models. It is not about integrating any API and bragging about innovation, but about training systems that understand the internal reality of the company. In detail. The company has developed an AI tool applied to product master data, the heart of its system. Automate the generation of information when new assortment arrives. And it detects inconsistencies that could break the chain: from label errors to coding mismatches. Pajares defends that therein lies the competitive difference: “There are very advanced models to program applications, but none natively understands how software is developed within a specific company.” The key is that the AI knows and interprets Mercadona’s own context, in this case. Between the lines. The strategy has two speeds. For business processes, such as planning employee vacations that do not compromise operational continuity, the company trains its own models. For more standard tasks, such as automatically recognizing supplier invoices, use off-the-shelf generative AI, while maintaining flexibility to switch suppliers when appropriate. Yes, but. Mercadona does not want technological disorder. It has defined an internal strategy that standardizes developments, guarantees common practices in quality and security, and prevents each team from creating its own isolated solutions. “AI cannot grow in a disorderly manner,” insists Pajares. This last point is key: in many companies, AI has been assumed as an engine that each team executes independentlyoften with different tools between teams, complements disconnected from each other. Mercadona seeks that cohesion between departments. The background. Pajares is clear: it is not about having the best algorithm in the world, but about knowing where to apply it. “In technology we tend to fall in love with the algorithm; in real life, intelligence lies in knowing where to apply it,” according to The Spanish. Mercadona’s bet is to predict demand with more precision than anyone else. Four decades of data on purchasing behavior, product rotation and logistics efficiency are your moat particular. And AI amplifies that advantage. In Xataka | Mercadona grows, but the “shopkeeper” model is dead: Spain has lost 142,000 businesses in 10 years Featured image | Mercadona
