/

Rethinking the Fashion Industry with the Artificial Intelligence: neural networks to shape a predictive Supply Chain

 

People around the world are going through a global crisis which is afflicting also the Fashion Market. What used to be conventional, today is odd and the sector must be knowingly ready to change. People are facing a shakeup of their daily routine and the wind of change is also affecting the Fashion Retail players that must rethink their purchase patterns, reconsider the “store” concept and satisfy the need of new consumers.


For them the biggest concern is to shop safely, and the retailers must think how to serve them better. The report “The State of Fashion 2020”, published by McKinsey, emphasizes the clear metamorphosis of the consumers: far from being passive players, now they lead the purchase experience, they are increasingly more demanding and ask for a “problem-free” shopping journey, an unforgettable experience, possibly customizable and gratifying.  

In any case, the route change of the consumers should not be seen as a deal breaker. On the contrary, the retailers should consider it as an opportunity to detect the weaknesses of the products Supply Chain and to think about innovative business models based on customer-oriented strategies to run faster than the Fashion Industry. It is both a technological and cultural challenge which brings together all the Fashion Players led by common keywords: digital innovation and lead nurturing

 
 

Have you ever seen the perfect sneakers or handbag while window shopping, only to be disappointed when you enter the store and find they don't have your size or desired colour? Well, these events are more common than one can imagine. It’s a known truth that inaccurate tracking systems and poor visibility result in a series of inefficiencies within the Supply Chain. The effects of a sub-optimal Replenishment can disappoint the performance expectations and reduce the degree of satisfaction of loyal customers, putting the business at risk. 

 

How should we reconsider the concept of “store” in order to meet the needs and demands of the 21st-century consumers?

First of all, the Retailers should bear in mind that the biggest obstacle to reaching sales goals is the demand high variability, with the addition of the difficulties of managing seasonal/non-seasonal items. For this reason, a flexible system to re-assess goals and strategies should be dynamic and scalable.  

Being a recurring theme, the complexity of the Replenishment process are rather well-known by the Retailers. The most common issue emerging while debating with the experts of the sector is that related to the limits of sub-optimal allocation processes: this approach relies on the long-term experience of specialists who combine their expertise with a good amount of intuition in an attempt to anticipate the future demand. This method, that could be defined as “manual”, impacts several fronts: the shop shelves may be empty, the sales may decrease, the level of CO2 produces may raise due to adjustment replenishments and not least the customers’ expectations are disappointed. It is recommended to avoid this last situation, as it makes the Fashion Luxury Industry a unique and exclusive sector.  

The FAIRE project originates from these assumptions and the basic idea is that the problems of the Inventory Management must be solved abandoning algorithms bases on statistical formula and preferring a proactive methodology which implies real-time monitoring of the stock and higher visibility on the entire Supply Chain.  Hence the need to rely on technological solutions which have been supporting many IT fields for many years: the future is the Artificial Intelligence, also known as AI.  

From Siri/Google Assistant on our smartphone to the facial-recognition app, up to medicine or finance, the AI made its way in a wide variety of sectors and nowadays it is not frowned upon anymore. Instead, it has a considerable impact on modern industrial processes to the point that many companies are reaching a remarkable competitive advantage on the market.   Considering the Fashion Industry, introducing AI solutions means to look to future scenarios in which human choices turn into data to analyze and consider in forecasting the consumers’ next moves. 

 
 

The FAIRE project aims at creating a digital innovation HUB in the Fashion&Design sector where companies can interact and suggest technological solutions able to offer new experiences to the final consumer and to increase the efficiency of the product life cycle internal and external processes.  
As partner of the FAIRE project, we are making a contribution to introduce Supply Chain Planning techniques based on one or more training cycles: to do so, we focused on the most performing area of AI, that is Machine Learning, the sector involved in the management of mutable algorithms alterable through the experience. As it happens for human beings, the system is able to “learn”, shifting results and strategies in relation to the experience, the context and teachings learned over time.   The typical patterns of Machine Learning allow us to shape mathematical models similar to conceptual tools with the purpose of defining the structure of each object employed for the solution and the item features, including meta data. This last represents a sort of calling card for the products, as it contains visual features and composition information such as color, material, shape, variations and further details. Providing this information is crucial when comparing data during the process.  
Together with our project partner DSTech, we are modeling neuronal networks which can convey to Stealth® The Fashion Platform previsions on optimal Replenishment for single store or data, after receiving training data to process. Future training and potential modification in their model improve the networks level of accuracy.  We focus exactly on the predictive ability of Machine Learning to realize the effect described by Wayne Gretzky as follows:

 

“Skate to where the puck is going, not where it has been”

 

This means that looking back to the past is useful only in an effort to forecast the future. Our long-time experience can help us staying one step ahead of the sales changes.  Five months after the beginning of the project, we are now working on the initial steps of the solution definition. We decided to opt for an open architecture in order to facilitate the integration of Stealth® The Fashion Platform with AI digital systems of the present and of the future, avoiding any limits for our customers when choosing the most suitable IT solutions. As soon as we get the first results, we hope to provide the Store manager with an optimal prevision of Replenishment to optimize the company planning processes and to align demand and offer as accurately as possible.  

Replenishment: how and when? Stealth® supports its customers with new developments and prepare them to face unexpected events and mutable scenarios with efficiency. Human beings thinking ability and intellect can’t be never replaced, but AI and Machine Learning can facilitate success opportunities and improve the results both in quantity and in quality. The challenge is to listen to the consumers’ voices to provide targeted and efficient solutions. Maybe one day when people will go shopping at one of our clients’ shop, they’ll see the desired product is available in store thanks to our neuronal networks.

 

Giovanni Botta Architecture Lead of Deda Stealth