Sign up

Il Sole 24 Ore

Evo, the artificial intelligence from Turin that wants to compete with Amazon

This software, developed in the incubator at the Polytechnic University of Turin, examines fashion choices and uses that data to anticipate market demand. So far, around 20 companies have benefited from its predictions. And their ambitions are growing.

Turin, October 12, 2019

by Natascia Ronchetti

The first to believe in Evo was the Miroglio Group, which encompasses eleven brands, including Elena Mirò, Oltre, Motivi, and Caractere. The Alba (Cuneo)-based group wanted to understand how to better design collections, manage purchases, and plan production in the face of demand volatility. Best practices developed with Miroglio became a case study at the Harvard Business School in the US.

Since then, Evo and its software built using artificial intelligence have helped about twenty fashion houses make the right decisions. For example, Evo can help brands understand why the market for the fashion industry in Milan is very different from the ones in Bologna or Paris: if a t-shirt sells well in Milan, it does not always mean it will have the same success in Bologna or Paris. Most importantly, Evo has learned how to anticipate demand trends using predictive analysis.

Evo is a young company. It grew from the incubator at the Polytechnic University of Turin, and today it has two offices— one in Turin, another in London— and 35 team members, almost all of whom are data scientists. “Predicting overall demand, ultimately, is not difficult”, says Fabrizio Fantini, Evo’s Founder. “But if we limit ourselves to this, we do nothing but follow the parable of Trilussa’s half chicken— an old story about statistical averages. Consumers have changed; they no longer buy what they find, but rather buy what they want”.

Evo uses its algorithms to break down products into data. It transforms products into “functional attributes” (such as the length of the sleeve, the type of neckline, and so on) and “style attributes”, which concern, of course, the style of the piece, using insights from a database that mapped behaviours of 1.2 billion consumers. Then the software “metabolises” relevant market data selected based on client attributes. Prices, products, etc., as well as the characteristics of the client’s main competitors. In the end, Evo transforms raw data into predictive markers of market trends so that clients can make systematic and profitable decisions.

Fantini is a management engineer with a doctorate in applied mathematics and an MBA from Harvard University. He focused on the world of fashion, he explains, when “I discovered that for companies operating in the fashion sector, inventory management is one of the main problems”. Today, the system has developed such that it can support fashion houses in production planning, stock allocation choices, and pricing and promotional decisions. It does this by combining advanced machine learning methods with statistics by taking into consideration many variables—everything from historical sales to the geographical area, and even the climate.

“Our dream now,” says Fantini “is to allow companies to compete with global giants like Amazon that have huge amounts of data available”. Not just in the fashion industry, where Evo’s core clientele comes from, however. The next challenge will be applying these practices with large-scale retailers to avoid food waste.

Hey! Was this page helpful?

We’re always looking to make our docs better, please let us know if you have any
suggestions or advice about what’s working and what’s not!

Send Feedback