Predictive algorithms are fully deployed by the government. For detection, to assess care needs or to predict whether a student will drop out of school. But what about business? And do you notice anything about it?

Predictive algorithms are computer systems that look for connections in large mountains of information, in order to make predictions and even decisions. Those systems are everywhere, explains professor of artificial intelligence Maarten de Rijke of the University of Amsterdam. "Make no mistake: banks, healthcare institutions, services and also farmers use them. Not only large tech companies."

Before we go into it further: how does such a predictive algorithm actually work? We explain that below:

Back to the algorithms in practice. Banks use predictive fraud detection algorithms: "There are tens of millions of transactions per day," says professor De Rijke. "You can't save that by hand, algorithms are looking for transactions that smell bad."

And farmers also use them. "They have become my right hand," says dairy farmer Wilfried de Bruin. "Sensors record everything and predictive algorithms recognize the deviations in walking and lying behavior of the cows. I get a message about that, it saves me work and time. And my cows have also been better since I have those systems. I have had much earlier a sick cow in the picture. "

A more famous example of a company that uses predictive algorithms: That company uses hundreds, if not thousands, of algorithms to determine what you see with them. "We use algorithms and machine learning models across the entire platform," says a spokesperson. "Before you get to the site, our algorithms start working. They make a prediction where you want to land when you come to the site and what offers you want to see."

Those algorithms do that based on information about you and the behavior of other users. "See the algorithms as a living organism. Even if we know nothing about you, we can give you recommendations based on your location. We know what Dutch travelers are looking for. We show you. But what Dutch travelers are looking for constantly changes, so the continue to learn algorithms and the range keeps changing. "

Professor De Rijke will continue. "Algorithms are everywhere. Web shops and newspaper sites use them to make recommendations from their online offering. Hospitals use algorithms for planning, TomTom to make maps better and KLM for aircraft maintenance," says De Rijke.


The Netherlands is on average ahead of the implementation of predictive algorithms in Europe, according to a recent survey by Microsoft and consulting firm EY among 277 European companies. EY and Microsoft spoke to nineteen large Dutch companies for their research. They use algorithms for a variety of tasks.

What is striking in the research is that more than half of the companies surveyed are still practicing with predictive algorithms: for example, they have pilot projects. "We set up pilots in the Netherlands quite easily to see if it works within the organization," explains Microsoft's National Technology Officer Hans Bos, who was involved in the investigation.

And algorithm specialist Frank van Praat from consultancy firm KPMG also sees this. He is involved with artificial intelligence and is often visited by large companies. "There is a dichotomy. On the one hand you have companies that are built on predictive algorithms from the start. Companies like, TomTom and Thuisbezorgd. There it is omnipresent. Look at more traditional companies, such as insurers and banks then the use of predictive algorithms is also regularly in the test corner. "

De Rijke adds that the second category is hard at work catching up. "Traditional companies must also become tech companies, otherwise they will no longer exist in a few years." And they do that too. "Banks, insurance companies, newspapers. You also clearly have parties that do nothing, but machine learning is everywhere. It is not something of the future, it is already there," says De Rijke.

SMEs are lagging behind

Only one in five SMEs in the SME sector currently uses information that they have or can collect about their customers. That is what MKB-Nederland says and according to studies by the Chamber of Commerce and by Kantar TNO and ABN AMRO. "That is more than the European average. But there is certainly room for improvement, because machine learning offers significant opportunities," says Feroz Amirkhan of MKB-Nederland.

"You have the front runners," says spokesman Ursul Schaap of the Chamber of Commerce. "But in general, companies are just starting to use data. And the general rule is: use it with common sense. It's not even about predictive algorithms or machine learning."

Amirkhan from MKB-Nederland adds: "Our view is that small and medium-sized businesses still do little with big data and predictive algorithms across the board, while there are great opportunities in converting data into knowledge."

The business community throughout the Netherlands therefore works to some extent with predictive algorithms or is about to do so. Handy of course. But there are also objections. For example, there is a risk of discrimination when using algorithms. And it is often not clear whether an algorithm is used. Aleid Wolfsen, chairman of the Dutch Data Protection Authority, recently warned that the government should be more transparent about how your data is processed and how decisions are made.

And even with companies it is not always clear whether they use algorithms. "Companies do not always find out that they use algorithms to help them make decisions," says KPMG's Frank van Praat. "They are not always transparent about it. That should change." Recent research by KPMG shows that few people are aware that algorithms are used. And that confidence in algorithms is low.

Load More Related Articles
Load More By WeeklyNews staff
Load More In Business

Leave a Reply

Your email address will not be published.

Check Also

White Ferrari Enzo turns up in one-off sale

Ferrari originally wanted to offer the Enzo only in red, but added the shades of black and…