In a room in Madrid, an avatar in the form of a nymph greets attendees and asks them what they will have for dinner this week. It is not a video game demonstration, but Sabia, a conversational supermarket based on artificial intelligence that proposes an idea that until recently seemed improbable: shopping by speaking, without browsing a catalog or comparing dozens of references. The scene seems futuristic, but it may be too early to talk about the supermarket of the future. Sabia is, above all, a sign of where the distribution is looking, and it is not just looking at the conversation.
Spanish large-scale distribution is incorporating artificial intelligence to optimize the supply chain, manage stock, combat waste, personalize customer relationships, and automate tasks. The question is whether any of these applications will actually change the way people shop and, above all, if the change will make economic sense.
According to Pablo de la Rica, Retail and Foodservice Knowledge Manager at Aecoc, the business association for large-scale distribution, the answer lies in understanding how the consumer has already changed. "We are facing a phygital consumer who does not think and differentiate so much by channels but by consumption moments and the most convenient options at each moment," he explains. The buyer has been connected for more than three decades, he points out, which has made them much more informed, capable of searching, comparing, and giving opinions before, during, and after each purchase, forcing brands and chains to bet on transparent and valuable information if they want to earn their trust. This does not mean, De la Rica insists, that the physical store is doomed to disappear, but that it transforms into a complementary space, with a more relational, exploratory, and experiential role, almost socializing. What has changed, and very quickly, is the habit, in less than two years, consumers have started using AI recurrently even in something as everyday as grocery shopping.
Carrefour has already made AI one of the pillars of its digital transformation. Its ClubIA offers members product recommendations and recipes tailored to their habits, as well as personalized promotions. The company also uses generative assistants to manage products, logistics, and internal processes, with its own advisory board to accelerate the integration of this technology.
Alcampo provides a less spectacular but perhaps more revealing example. Its Smartway application uses AI to identify products close to expiration and propose actions such as reducing their price, and according to the company, in 2023 it prevented more than six million products from ending up in the trash. Here, artificial intelligence does not change the consumer experience, but it changes a bottom line.
The Valencian company Consum, on the other hand, is working on hyper-personalizing the customer, based on their consumption habits and how they communicate with each user. They are also exploring intelligent assistants to manage internal information and automate administrative tasks. Therefore, the sector is not waiting for the future to arrive; it is already testing where AI can be profitable.
Aecoc sums it up bluntly, "digitalization is not an option, it is an obligation," De la Rica points out, for whom hitting the mark on what the consumer is looking for, where, what, how, and when, no longer depends on having AI or not, but on how long it takes to apply it correctly. Technology, he says, has ceased to be a competitive advantage, as any technological advancement could have been decades ago, to become a disadvantage for those who decide not to consider it.
However, it remains to be seen whether all this movement is already profitable or still a promise. When asked if the investment in AI already translates into sales or loyalty, De la Rica responds cautiously, "we are seeing AI developments in a short time to be much faster, efficient, and effective in customer relations in attention, treatment, service, and loyalty," he explains, although he prefers to talk about a new layer of value that provides differentiation rather than a guaranteed economic return.
This caution makes sense when looking at the bottom line. The food distribution sector operates on tight margins, and a new technology has to justify its cost with increased sales, higher purchase frequency, loyalty, operational savings, or less waste. An outstanding experience is not enough. Additionally, AI can alter competition. For decades, supermarkets have perfected the way they organize their stores to influence our decisions, placing candies near the checkout, strategic products in highly visible locations, and on the internet, search became the new gateway to commerce. Now that gateway may change again. Instead of searching and comparing ten products, the consumer may simply explain what they want and accept a recommendation.
This concentrates power. If an AI recommends a product instead of showing ten, whoever controls that recommendation may end up controlling a significant part of the purchasing decision, and small brands could be even further from the consumer. But it can also happen the other way around, specialized AI can make products visible that are currently lost among thousands of references, understand that someone prioritizes certain ingredients, wants to buy locally, or is looking for a healthier alternative, and turn a massive catalog into a conversation. AI can, therefore, concentrate decision-making power or return it to the consumer, and we still do not know which path will prevail.
The evolution of retail itself advises caution. Amazon Go promised cashier-less stores and became one of the great symbols of future commerce, but then the company scaled back the model and closed a good portion of those establishments. The technology could work, the question was whether the economic model justified its cost, and the same question now applies to AI.
Perhaps in ten years, we will do our shopping by talking to a machine, or perhaps we will continue to enter the supermarket, look at the shelves, and compare products. What we do know is that major chains are already investing to find out. And that is the battle that begins now, not to build the most futuristic supermarket, but to discover who can make artificial intelligence generate enough value for consumers to want to use it and for the business to afford it.