Investing in AI is a necessity for businesses

Fontys School of ICT

Can business wait and see when it comes to AI?

After the Internet revolution, we are in a new era. Artificial intelligence (AI) has a transformative impact on many aspects of daily life, but also in business. Eric van Tol, director of the AI and Big Data expertise centre at Fontys, sees many opportunities for innovation in companies, especially in service. But also that applying AI and thinking about data is not a long-term issue. No, we are lagging behind and if the Dutch business community wants to maintain its strength, then investing in AI is a necessity.

International competitiveness

Eric van Tol has been saying for a number of years that we are lagging behind: "A professor told me 'we are extremely smart', and I have no doubt about that. But we cannot match the scale of data collection, budget and experts that China can bring in. This is starting to sink in in the Netherlands, and you see that we are organising ourselves nationally. But we are a gnat compared to a mammoth. This has to be done at the European level."

The biggest challenge in the Netherlands, and Europe, is expertise. Training people with hands-on knowledge of data science and AI applications. The traditional formula does not work here, according to Van Tol: "I once saw a curriculum that showed technology from four years ago. These are our future professionals, and you are going to send them into the market at a disadvantage. And the market wonders why they have to cough up the budget for this. HBO schools are perfectly capable of providing the knowledge that the business community needs, but cooperation is required to do so. But what should we focus on? Erik sees many opportunities in the use of data and more complex applications: "What we do with data now is often analysis. Business intelligence can tell you that you went bankrupt three months ago. What if you could do that differently and predict correctly? That's where we need to go."

Data-driven service innovation

If we want to move beyond traditional machine learning, we need to think differently about data, processes and collaboration. And then opportunities arise, van Tol realised after his time as the driving force behind the Dutch government's breakthrough project: "One of the things I learned from the breakthrough project is that we look for opportunities in data push applications too much. But there are many opportunities in data-driven service innovation. Data is the big driver and in companies you see that all the major innovations right now are in service. Look at the big guys like Facebook and Google, that's service with data."
In service innovation, van Tol sees a number of fixed characteristics: providing real-time insight, enabling predictability and making deviations visible immediately. "All companies know what attributes such a service innovation has. You can then think about that yourself as a company, which aspects you can do something with? Of course, you can offer such a service on the basis of analysis, but also as an experience environment. That too is data-driven innovation."

Unlocking untapped data

"Where we often apply AI is in small actions. Recognising cancer cells, trends on Twitter, you name it. But if you want to build more complex systems, you need to understand much better what data you have." Data visualisation is the key there. Van Tol gives the example of complex transaction data that is stored but not yet used. Dark data is what we call it, because it is not used for deeper insights. "If you then visualise it, a fraud expert can see patterns in it that you can use. So you can unlock data and get much more knowledge about your processes and also make predictions on that." Besides analysing data, you can also turn the story around according to van Tol, thinking from data. 'Natural occurring data' is another way of looking at data, he explains: "Suppose you want to analyse road behaviour, then you can start collecting data from drivers. But you can also use data from TomTom, because that is already there. That is an example of 'natural occurring data'. Another example is satellite analyses of parking spaces at shopping centres, to predict sales. Or similarly registering oil barrels, which would allow us to predict the total supply of oil fairly accurately. That's already possible and an example of data-driven service innovation." Data is there for the taking, and what is not there can be generated through gaming and simulations. Unlocking that data is also the key to meaningful innovation for business.

Eric van Tol talks about the future of AI and business in the second episode of the podcast AI Garage. He does this with Jan-Kees Buenen, CEO of SynerScope B.V. and host (and Fontys University ICT teacher) Erdinç Saçan. Listen to it here or on Spotify. ​