Sign up


Artificial intelligence and more

Turin, February 8, 2021

Interview by Sofia Bergami

Sofia: “Good morning from Podstem. This morning we are together with Fabrizio Fantini, founder of the Evo Pricing company. Good morning.”

Fabrizio: “Good morning.”

Sofia: “Everything good?”

Fabrizio: “Everything is fine, we are finally back out after the Christmas break.”

Sofia: “So, today we are going to talk about artificial intelligence and we will try to give a slightly more captivating approach, so we will talk a little about artificial intelligence as such but we will also discover a little bit the relationship that artificial intelligence and technology have on aspects of daily life and not of the human being. What is artificial intelligence?”

Fabrizio: “Artificial intelligence is in fact both a technology and a field of scientific research. Today with the public let’s focus on technology which is what is of greatest interest if we want for everyday use. This technology is used to do many things, but starting from the base, what is it? It is essentially a method of data analysis, which is an evolution of everything we all studied in school. At school we studied classical analysis techniques, so time series analysis for example, seasonality, rather than regression for those who remember and the concept of correlation. Artificial intelligence is a bit of an evolution of this, I call it a correlation, as the English say “on steroids”, so on an exponential level. That is, no longer just as in the classical world, a correlation between, for example, price and volume in the world of elasticity, sensitivity and price; or rather the climate and the number of umbrellas that are bought, the classics. Artificial intelligence is able to analyze much larger amounts of data and therefore have more input signals and try to understand multidimensional correlations. What does this mean? For example, innovative data such as images and videos, which are nothing more than different forms of data, are no longer just numbers but combinations such as text and images and in fact always have the same representation matrix. So these techniques are able to analyze these large amounts of data and see correlations, let’s say multidimensional and therefore they are technologies able to do better what we have already tried to do for many years. That is, for example, thirty years ago, twenty years ago there was talk of expert systems to try to model how medical doctors make diagnoses, and however it was discovered that in reality it is a very complicated process that of diagnosis and is not easily modeled with a neural network as it was called then. So these expert systems had lost a bit of sex appeal and in recent years as we all know, the cost of storing, of processing data has dropped exponentially and is much easier, much cheaper. So, for example, a complex process such as diagnosis is able to approximate but not replicate the human capacity, but it is possible to approximate some parts, the most repetitive and methodological with great success. It is clear that much of the knowledge about artificial intelligence is still in the making and much of what we sadly hear in the press every day is a bit misleading, in the sense that asking if artificial intelligence is better or worse than human intelligence it’s a bad question.”

Sofia: “Yes, this thing makes me smile, especially this question which is the first to begin, which is clearly the opening one because perhaps in the ignorance of those who do not deal directly with artificial intelligence, I take myself as an example, when one thinks of artificial intelligence, the first thing he thinks is trivially a robot, but artificial intelligence is not just a robot but, as you also specified, it is a very broad system of course. I really wanted to ask this question because I realized that, in a nice way, when a person thinks of artificial intelligence he imagines a thinking mind that is moving, instead it is quite the opposite.”

Fabrizio: “Absolutely, indeed it happened to me just this week in class at university, that a student asked me:” But is the robot that cleans your house a form of artificial intelligence? “. And so this example in my opinion helps to understand the phenomenon a bit, that is, what is artificial intelligence inside a robot? Typically some things, some elements of robots certainly pertain to artificial intelligence, such as the ability to recognize images, which is a skill that requires these evolved correlations. There are also many things that are classic technology techniques, such as the control of the movement of the robot are methods we say that existed even thirty years ago, there is almost nothing new. I would also like to specify in the analogy between the toaster and the blender, that the human brain is also able to make inferences, let’s call them top-down, that is to look at things and processes in a holistic way and therefore to understand what is happening and this is all the part chemistry, it’s all a more complex part of the brain. What these techniques do, these artificial intelligence neural networks, try to replicate the electrical part of the brain, it is only a part in reality, we do not yet fully understand how neurons in the brain work. And so in reality the machines are able to look only at roadmap things, that is, starting from the data and trying to pretend to be intelligent, that is, in reality what these machines do is replicate the patterns in the data on which they are formed. So what happens? That for example the robot is trained on the recognition of floors, so it seems very intelligent, but if you for example put it on the grass or on a surface on which it has not been trained it would not know where to go.”

Sofia: “In this regard, I would like to ask you: why is this inevitable relationship of man with technology so discussed?”

Fabrizio: “Well this is in my opinion a bit of a combination of two factors. The first is our general resistance to change which is nothing new, in reality the exceptionalism of the present always makes me smile, that is the tendency to think that today we are in an absolutely strange world, different and new compared to what it was 50, 100, 200, a thousand years ago. In reality there is nothing new if we think about how much the introduction of other innovations such as: electricity, the train, the locomotive, the typewriter or the washing machine. These innovations have resulted in gods very impactful changes. Just remember a hundred years ago more or less the reaction to the car: everyone riding a horse and a carriage and the car was taken as an absolutely frightening innovation. So, in my opinion, this is an aspect where there is nothing new. There is another area which is that of information, in reality many people still do not have a thorough knowledge of what is actually possible and what is not possible. Unfortunately, they are informed by articles or press coverage that are clearly looking for news and sensationalism and therefore talk so much about the singularity, the famous event in which the machine will overtake man. Today there is not even the theory to get there, that is, it is not just a problem of computation capacity, cheap, so to have a bigger computer, we just lack theory, so it’s science fiction. It is not said that in a hundred years or a thousand years it will not be possible, just as it is not said that in a hundred years, a thousand years it will not be possible to go to other galaxies. Here we really get into science fiction. I think part of this fear is absolutely unjustified.”

Sofia: “I would also like to say that clearly now in the world situation we are also very attentive to the planet Earth, which is why I ask you: is the relationship between man-technology-eco-sustainability also important? And above all, if it has started to be so important in recent years, when everyone has clearly started to be very interested in eco-sustainability as well.”

Fabrizio: “But it’s not really a compromise, it’s what I would technically call an objective function, that is, the technology can be used to do things depending on who creates it and actually the AI ​​technology is very effective in removing waste. And here before talking about examples of artificial intelligence, I would also like to give a simple economic example: even if the theory existed to replicate the human brain with a machine, today it would be absolutely economically suicidal. The level of energy that the human brain requires is about 15 times a few grams of sugar, that is, extremely efficient in its use and replicating it with a machine would perhaps require megawatts, gigawatts of power and then space, construction cost, etc. So precisely economically in this man-machine relationship there are two different efficiencies and the efficiency that the machine can bring to the world is truly enormous, being systematic and on some things the machine requires much less effort than man. Let me give an example that is being talked about a lot these days: the self-driving car. The self-driving car actually allows a considerable saving of energy. Because? Point one: we buy cars that 95% of the time stand still in a parking lot, garage or some place where they are not being used and clearly with a self-driving car, a model where ownership is shared and ‘use of the car becomes much more efficient requires to produce far fewer cars, which means less waste of resources, therefore less waste of metals, fabrics, etc. and less land consumption. Why how many urban spaces require all these parking spaces on all sides? And therefore it is truly a transformative potential and very similar to what is possible in many other areas, such as production waste that can be significantly reduced. Or like relocation, one of the reasons why it is produced in China is that it costs less, but it takes much longer to transport the material to Europe or Italy. So thanks to artificial intelligence being able to observe better, more accurately the demand it is also possible to bring production back to the premises because there is more certainty of the purchase and therefore there is greater convenience, even perhaps to invest a little more in production. In reality they are transformative technologies that some things do better than human beings and therefore in terms of sustainability they really manage to help us achieve these goals.”

Sofia: “Why is it that if on the one hand there is the great demand, on the other hand we have had this great difficulty in welcoming technology and innovation when it was needed and at a time when it would surely have given a very big hand?”

Fabrizio: “This goes back a bit to the point we made earlier about the fact that resistance to innovation is nothing new, it is a fundamental part of being human because unfortunately we are fascinating creatures in terms of the ability to adapt and evolve, but let’s remember that let’s say the brain’s goal is to save sugar. The brain is a very lazy machine and it is evolutionarily convenient to have been the design of a lazy machine, in the sense that we must remember that until a few thousand years ago it was very difficult to find food and therefore the brain was trying to save energy and not to make big efforts. Change requires effort, even the easiest change, the most convenient, most profitable change, however, requires effort because changing a habit inevitably requires reflection, requires you to take a risk, and inevitably requires you to activate the higher functions of the brain. So in reality it is not a surprise, indeed last year there was this joke that asked companies a bit: what was the strongest spring for you for digital transformation? One of the answers was Covid, it was neither the profits nor the market nor your boss … then when you are forced to change, unfortunately it is clear that you always react a little late to events. But this is also a bit rational, in the sense that how many changes may actually not work and therefore if we always wanted to try everything new, we would inevitably scatter and we would no longer be able to do anything productive, so there is. it is always a tension between the new and what remains. There are some things that have happened, which is good that it turned out. For example, we have always had a team distributed worldwide, so we have always known smart working that it was a resource because it has always allowed us, even as a child when we were only 4-5 people, to have access to different talents to which otherwise we would never have been able to aspire locally only with the people in our small neighborhood. It is also true that the excess of smart working can also lead to a pulverization of the richness of human relationships, like all innovations, as well as e-commerce. It is true that e-commerce is a great thing because you can comfortably have everything you want at home, but be careful that even there if one stops and thinks, in a world where everything you need arrives at home, the contents videos arrive at home and therefore you no longer go to the cinema, you do work from home, like saying you are born and die at home.”

Sofia: “All very alienating.”

Fabrizio: “So it’s a world we want to live in? Returning to the theme of eco-sustainability, for example Amazon is not so positive for the environment because the classic store was much more efficient from a logistical point of view and therefore consumed both less CO2 for transport and less packaging for what concerns the product. So a somewhat alienating world in which they don’t exist more shops, the office no longer exists, the traditional community no longer exists is not necessarily a good, we need to find a balance.”

Sofia: “Exactly, absolutely. But if this thing made me smile a little bit that we are always looking for new things and when it was precisely the time to get involved it was difficult to actually do it, for x reasons. So yes, it is certainly true that adapting, the idea of ​​investing certain resources in any case can also be a little scary, but it was a point that I wanted to touch because on the one hand it seemed a little bit of a paradox to me, because we always expect anyway, when we have a way to do it we withdraw or at least we are reluctant.”

Fabrizio: “However, this is one of the things that, if we also want to technically and therefore also mathematically, differentiate man from machine. There are also mathematical tests to say when the machine becomes truly comparable to man, it is the ability to synthesize paradoxes and certainly today it is something that differentiates us. No matter how complex the instructions of a machine, they are always algorithms, therefore replicable sequences. While, on the other hand, we humans are able to manage this type of problem in which in reality the world is not made up of black and white colours but is made up of many paradoxes and many compromises. Like so many things, reality is in the middle so smart working is a positive and a negative and therefore it is a paradox in a certain sense, but there is a balance in which it becomes desirable.”

Sofia: “How then, to date, innovation and technology can meet the needs of companies, but not only of companies, also of customers?”

Fabrizio: “Today we discover that there is innovation at the company level at three levels, so it is not just one type of innovation, but there is innovation both of product and of process and also of management at all three levels. Of product, in the sense that it is perhaps the easiest thing to visualize and understand, there are new products. For example, the self-driving car we were talking about before is one of the new products that is being finalized to be launched and so are many other new products made possible by technology. There are process innovations, therefore new ways of doing things that were already done before, for example, but it is nothing new. For example, let’s think of the turnstiles for entering the company, once you entered with the guard who controlled, then the badge was inserted and now maybe there is it is facial recognition or even the chip under the skin, which is a bit scary but in reality maybe a way will be found even without any implant in the human body. So process innovations, the same thing done in new ways, as well as in the factory there are so many innovations born on quality control, preventive maintenance, etcetera etc etcetera. And then there is a level of managerial innovation in which the way companies are managed is also changing. Some of the things that were historically done by hand and made by men, today machines do better. But why? Because in reality the demand, both that of us consumers and that of companies, becomes increasingly complex to satisfy, there is more competition, there is more volatility, there is it is more geographic differentiation and therefore it is very difficult for a manager to keep up with customer demand. This, on the other hand, is one of the things that the machine does best, because it’s like having a battalion of analysts, having a semi-intelligent machine because then I go back to saying, this artificial intelligence is a very limited meaning, not even adolescent. This type of techniques allow you to better understand in real time at a probabilistic level, mathematical and to understand what various people, various sections of your population want in real time. And then they compare the manager with the traditional dilemma of choosing me because I believe based on my experience that it is right, compared to, he chooses the consumer because through his data, then through the revealed behavior, I discover or teach me that he wants something different from what he wanted until yesterday. This tension means that the self-centered manager will have less and less space, more and more will be needed a quantitative manager capable of accepting that in part he must create a story and in part he must accept the perception of the reality of the market.”

Sofia: “Ok, let’s talk now instead of the Evo Pricing company. What do you offer to customers? What are your prospects? How was it born?”

Fabrizio: “Let’s start with how it was born. It was born six years ago from my PhD experience in applied mathematics and ten years of consulting at McKinsey. Putting these two baggage of experience together I saw that there was literally a great disconnect between what was possible to do with data and what was actually done even very advanced like McKinsey. So I tried to implement my underlying intuition, which is a very simple intuition, that is, when the world becomes complex and unpredictable, explicitly trying to predict and decide based on prediction is inefficient. Evo Pricing is a bit if we want to at the frontier of what we call “dynamic discovery”, that is, continuously helping companies to keep up with the reality of the market. We do this in three areas specifically: originally in pricing, which is where the company name came from, then help price pricing for new products or update pricing and discounts for existing products more efficiently. But also in two other sectors: the supply chain and therefore we say all the decisions relating to inventory, purchases, warehouse management, transport, logistics and stores, wherever they are. And in marketing, then calculate the expected value of customers, the risk of following the customer, the financial world, rather than investments in customer acquisition and renewal. These three domains are in fact our scope of application, where we have a team of 43 people divided between London and Turin who deal with analyzing, developing software and then offering cloud solutions to subscriptions to services “software as a service”, therefore software services for companies that until yesterday had to be from 100 million upwards but from April of this year onwards we will try to increase even more.”

Sofia: “Very well. Thank you very much Fabrizio for this chat. Thank you very much for all your answers, for your availability.”

Fabrizio: “My pleasure.”

Sofia: “A greeting.”

Fabrizio: “Always a pleasure, if in the future you have any kind of questions, even short ones or for opinion, feel free to write me whenever you want.”

Sofia: “Very gladly, I’ll keep that in mind.”

Transcript: English (auto-translated) / Italiano

For better quality translations please reach out.


For press-related queries please contact

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