Flour, wheat, water, salt … and algorithms. Bread has also been transformed. Artificial intelligence is added as one more ingredient in its production process. The craftsmanship of a good sourdough has found in data science an ally capable of digitizing the tradition that surrounds the bakeries. Like any other company, the bakery needs to optimize the orders to the suppliers, the daily production and even the knowledge of the consumers. And here technology, strange as it may seem to find it in a loaf, plays an indispensable role.
Pan Delirio, in Madrid, has opted for this innovation. As Javier Cocheteux Jr., one of its founders, explains, his personal intuition to adjust supply and demand was failing more than he would like. “A few days there were 70 kilos left over. Others of us ran out of bread at noon and didn’t understand why. The only thing that was clear to me is that something was happening ”, he says.
After giving it a few laps with his father, the other founder of the bakery and with whom he shares a first and last name, artificial intelligence seemed a reasonable solution to so many imbalances. But it was only the beginning of a journey they took a year ago with the help of Olocip software. They had to supply a huge amount of historical data to the machine. From the most basic about production and sales to some peculiar ones like the daily weather conditions. Without debugging the information, the algorithms would never work properly. “There are many parameters that influence the purchase of bread. For example, Mondays are when it is sold the most and Fridays just the opposite; but there are more factors, such as whether it is a holiday or it is raining, ”says Cocheteux Sr.
For a workshop like yours, prediction, they say, defines the border between success and failure. Bread takes about three days to make, requiring that same time in advance to determine the total amount of ingredients they need. And their business is not only limited to bread, but they also produce pastries and roscones at Christmas. That is, they work with perishable materials. If they do not give them an almost immediate exit, they would only accumulate losses. “Artificial intelligence gives us information to assess what to do, although the final decision will always be up to us. The software It helps us make better decisions, ”reason Cocheteux Jr.
So far the theory because the pandemic has hindered innovation. When operating with data, the reality of the last year differs from the normality before the coronavirus. The consumer goes down less to buy, prefers other types of products, spends in a different way and thus a long etcetera that artificial intelligence has had to process. Despite everything, at Pan Delirio they have not backed down. They hope to have the algorithms at full capacity in the coming weeks. “There is fear or reluctance in more traditional companies to apply this type of tools, but the techniques and methodologies are useful for giants with 600,000 employees as well as for a factory”, suggests Asier Rodríguez, one of the data science managers at Olocip.
To lose the fear
Rodríguez comments that fear contains a good number of small businesses from betting on artificial intelligence. Obviously, your company is dedicated to this technology – you’d be surprised to say otherwise. He understands that reluctance stems from a widespread trend: to view algorithms as substitutes for flesh and blood workers. He is categorical: “Data science does not replace anyone. It is a complement”. With Pan Delirio he has found an example of what the qualities of this tool are. His daring, in his opinion, is the way to lose fear and realize how useful it is.
These little dabbles with intelligent machines only represent the beginning of a more complex path. In the workshop they want to go further with the transformation. Cocheteux father is not satisfied with fine-tuning the forecasts. He wants to dare with Industry 4.0. “The machinery has been stagnant for 120 years. The mixers are practically identical to those that were used two centuries ago ”, he laments. This reality means that, for example, to turn tons of dough it has to be done manually in small quantities or at once with a machine that rotates the container. “We have to give a major turnaround to baking in every way,” ditch.
We will have to wait a few weeks to find out if the arrival of artificial intelligence fulfills the purposes. At the moment, they only have the theoretical tests that the algorithm has developed. The results are promising, but daily activity will put everyone in their place. While doubts are being cleared up, traditional bread can already affirm that it does not conflict with the innovative avant-garde. “It is striking to complement the artisan with something totally technological. We are not afraid to install these types of solutions. It’s a bet. Joining both fields seems very nice to me ”, concludes Cocheteux Jr.
Eddie is an Australian news reporter with over 9 years in the industry and has published on Forbes and tech crunch.