To define Machine Learning very simply, one can refer to Arthur Samuel, pioneer in the field of computer science, who already spoke about the concept as early as 1959. According to M. Samuel, machine learning is the subfield of computer science that gives computers the ability to learn without being explicitly programmed.

In the last months and even recent years, Machine learning knows a tremendous amount of development and attracts the attention of many players of the retail industry. Now as a curious mind, I wanted to investigate to understand this hype amongst big players and get an idea of where this all leads. Therefore, as a retail aficionado and a Belgian I spoke to several Belgian companies* active in the field; and here are the different key insights I gathered.

How come the (sudden) hype with Artificial Intelligence?

You might wonder why I suddenly speak of Artificial Intelligence now, when I announced that I was looking into Machine Learning. Well don’t get too confused with the different terms. Intelligence displayed by machines or artificial Intelligence, is just a more marketing term to talk about machine learning.

As mentioned, machine learning is not a new concept, but it seems that the speed of development has increased only just recently. Well this can be explained thanks to big Data and the development of the cloud. Retailers, brand manufacturers and other key players can collect huge amount of data thanks to interactions with customers and enrich the data with other parameters (location, timing, ….). The access and the sharing of all this data can go through the cloud, which didn’t exist until recent years.


What is artificial intelligence good for?

Well the most important thing to remember is that AI can be used in many varied domains. In retail (because this is what interests us the most here), we can detect two major applications of AI: Prediction and Recommendation.

The amount of data collected about a consumer is put in the machine and the models are there to analyze his behavior and also predict his next moves. The concept of AI goes even beyond simple prediction as some analysts told me that it is possible to predict the behavior of consumers and also understand the ‘why’ of this behavior.

For recommendations, let me of course give you an example you all know. I guess many of you went shopping online on websites and received different items recommended according to your preferences or previous purchases. But in terms of recommendations you can also expect more than just that. Imagine your food retailer is looking through your data (actually the machine will do it for him) and sees that you always take your car to go run after work. He will know also your location, your preferred supermarket, and even your route. Therefore the retailer will be able to send you after your run and at the moment you are in the proximity of the store a more targeted offer or add for a sport’s drink for example.

To say it in simple terms, Artificial Intelligence is a tool that will boost personalization.


Now all eyes are on AI, it will develop fast no?

During the different interviews I conducted, all specialists, and agencies could detect a few roadblocks that need to be addressed.

To start, it is difficult to find the right people with the right knowledge. There aren’t many specialists existing in the field and the freshly graduates are too often formatted in one way or another. Either you find programmers without mathematical knowledge, or you find math experts without enough programming skills. In addition to that, Belgium experiences the problem of high taxation. If one wants to attract outsiders to work on projects or educate the young about AI, the packages are almost unaffordable for universities or companies due to the Belgian fiscal constraints.

To continue, the access to data in Belgium is not always easy. Surely the cloud allows many things, and companies have a lot of internal data. However, if you want certain data such as the state of the roads, or the weather in a very accurate way, you can’t expect it from national institutions and are still more advised to use private channels such as Ways owned by Google for example.

Finally, to control the budget while developing Machine Learning projects is difficult. This roadblock seems to be shared by all players and agencies over the world apparently. It is difficult to find similar projects to evaluate, and many developments are made with trial and error. Therefore, sticking to a planned budget is not always easy and you need a certain financial backbone. This probably explains why many of the advancements are studied and made by major names such as Amazon, IBM, Google or Alibaba just to name a few.


Will AI mean fewer jobs?

Now there is a question that some people start to ask. If machines do more, and can learn, and never make mistakes, does it mean that jobs will disappear? Well the answers are mitigated on this one. It is undeniable that thanks to machine learning some projects that used to take months or weeks can be done in a few days. Also, AI allows robots to replace lower skilled functions or repetitive tasks with low added value. So it could seem that people with lower skills will be in trouble. Think of delivery jobs, drivers of all kind replaced by driverless vehicles and so on.

However, the future is not so gloomy according to some experts. And this is for two main reasons apparently. First, creativity is still only a human thing. Creative robots don’t exist. Secondly, the interactions and the developments are mainly aimed today to have collaborations between online and offline at the service of humans. Some specialists predict that machines will help humans, so humans can concentrate on other tasks with higher added value.

The debate is on, but whether it is dark or hopeful, I’ll let you decide.


So how is the future for retail?

There are major disruptions and development coming up in the future. That is certain according to all experts I spoke to. The drive and search for optimization and better solutions will definitely be catalysts for machine learning in retail.

But AI is just a tool on top of all the tools that our industry has at its disposal.

One thing seems to be clear though; the road might be long, a bit uncertain, but humans will still have the lead role.

* Some of these companies I interviewed will exhibit on Paris Retail Week, so make sure to visit the different Belgian exhibitors. 


Between 19 and 21 September 2017, Paris Retail Week will once again showcase French Expertise and will be the only event in Europe which addresses the omni-channel problems associated with 360° retail.
This third edition will bring together in a single event E-Commerce Paris, Europe’s biggest cross-channel event and Digital(in)Store, the leading trade show devoted to the digitisation of points of sale and distribution.

Take advantage of this event dedicated to 360° retail uniting the entire community of on- and offline retail for an optimal vision of current retail issues!
To register as a visitor click here 
Free online registration until September 18th, 2017
Registration at the event: 50 Euros incl. VAT