AI, data science to find the right answers

I first heard this story in a speech by Paul Rulkens, a very high-profile consultant and popular keynote speaker. So if it sounds very brilliant to you, it is simply because I copied it. But it is a little story that I usually tell my managerial and entrepreneurial clients, when they are about to start an AI project, because it turns out to be very effective. In 1942, Albert Einstein was teaching at Oxford University. One day, he gave a physics assignment to the final year students. As he was walking around the campus with his assistant, the assistant suddenly looked at him and said, “Professor, isn’t the paper you just gave to the final year physics students the exact same paper you gave in the same class last year?” “Yes, yes,” said Albert Einstein, “It’s exactly the same.” “But, Professor, how is that possible?” said the assistant. “Well,” said Einstein, “the answers have changed.” The answers have changed. I find this observation brilliant and versatile, an effective closure to a very useful anecdote to stir the consciences of our most reluctant interlocutors to innovation.

AI, the answers change with the context

Because ‘the answers have changed’ what does it mean in practical terms for our customers, managers, entrepreneurs who have achieved so much so far? It means that everything that has led them to grow their business so far may not automatically lead them to the next step. And if you want results you have never achieved before, you have to start doing things you have never done before. Yet, in Italy, there is a hard core of managers and companies that still find it very difficult to change. Rulkens explains that when faced with an insurmountable problem, people, in everyday life as well as in their work decisions, generally adopt one of two behaviours: they continue to do the same things to a greater extent or they continue to do the same things to a lesser extent. Only very rarely, however, are they willing to do completely different things. When we think and act in this way, we are merely acting within patterns that we could hypothetically represent as a box. The Anglo-Saxons speak for this of ‘thinking out of the box’, referring precisely to this box.

AI, the Premoneo approach

At Premoneo, dealing with pricing, forecasting and segmentation, we know that the boundaries of the box are moral, legal, technological. Sometimes all it takes is to break just one of the walls of the box, the one made up of technological boundaries, to change the status quo and start a path of innovation within the company. Breaking out of the box also requires a certain amount of humility, to overcome the classic ‘we have always done it this way’ rather than ‘we already do the best’. One example concerns geo-pricing. Some time ago a customer explained to me, with a certain bravado, that he charged different prices based on the geographical location of his outlets, where a customer’s willingness to pay might be different, and then linked certain price changes to the average per capita income. They were surprised when, as a result of our analysis, they found that, for example, in a particular town in Abruzzo, for four months of the year, the pattern describing the behaviour of their customers was much more similar to that of customers in a large city in the north.

AI, analysing data to interpret and predict demand

So I decided to phone the store manager to ask if he had any thoughts on the subject. ‘Of course,’ he replied, ‘here since May we have many more customers, pensioners from the north, on holiday. And unfortunately for my generation, obviously those pensioners have a higher per capita income and propensity to spend than the residents. Did one have to think outside the box to come to this conclusion? Yes, or perhaps it would have sufficed to rely on Artificial Intelligence to analyse vast masses of data, to unearth patterns, so to speak, invisible to the human eye, especially when plagued by the fatigue of change and the presumption of having already tackled the problem with a correct strategy. Fortunately, according to a reliable study by Horvath, 79% of companies plan to invest in AI by 2030, for price optimisation. A clever way to remind us that the answers have changed.



Leave a Reply

Your email address will not be published.