Using Artificial Intelligence to Balance Out Customer Value

December 13 it was that time again: the second edition of #projectwaalhalla: Social Sciences for Startups. This time with Peter van der Putten speaking as data scientist. He is guest researcher at the Data Mining Group (algorithms research cluster) of the Leiden Institute of Advanced Computer Science. He is also director of decisioning solutions worldwide at Pegasystems. There is, according to Peter, a lot of potential for new startups in this area. Are you going to be the next success story ?

I am actually very curious what you, as a leading data scientist, think of this whole big data thingy.

I am fascinated by learning from data, but have mixed feelings about big data. The concept is being hyped a lot at this moment, while the algorithms to learn from data have been studied since the 40s in computer science. Many of the “modern” big data technologies like Hadoop are in fact still limited frameworks for old-fashioned, offline batch processed data, instead of real-time processed data. The focus should really not have to be on the data, but on the analysis — how we generate knowledge and learn from data through data mining — and more importantly, how do we operationalize this knowledge, how can we use this knowledge. Because: “Knowledge is not power, action is.”

And what is the role of psychology in big data? And of philosophy?

Psychology has begun studying intelligence fifty or sixty years before computer science did. People, animals, plants and all intelligent systems are basically information processing machinery. Psychology seeks to understand these systems and to tries explain behavior — if you understand a bit of that system, you can use this knowledge. For example, to teach computers, stupid mathematical pieces of scrap, smarter functions such as learning and responding to customer behavior. What is to say, thinking the other away around, that people don’t have to think like computers. See for example the psychologist Daniel Kahneman who won the 2002 The Sveriges Riksbank Prize in Economic Sciences, the unofficial Nobel Prize in economics for his insight that people aren’t rational agents that properly weigh all the choices before deciding something.

And philosophy? These guys have dealt with big data for more than 3,000 years now. Just think of the nature vs. nurture debate: do we acquire intelligence and other properties by (data) experience or are they innate? Or the whole philosophy of mind discussion, with roots in the ancient Greeks: what do we really know? And is there is only experience or just reality?

Do you like what you have read so far? Get a quarterly update of what I am busy with.

You have a background in artificial intelligence (AI) and even studied with the famous and wildly attractive Bas Haring (Not related… well cousin to be honest. If you insist). What could AI mean for business, and how is it different from Big Data?

As long as AI is not used for old fashioned data manipulation or poor reporting, but really as intelligent data science, big data is one of the tools within ‘learning’ artificial intelligence. That is, systems that are not smart by the knowledge that is pre-inserted, but which have the capacity to learn and combine what is learned with background knowledge to deduct decisions. This is what I like to call the field of ‘decisioning’. Really intelligent systems put that knowledge into action and are part of an ecosystem, an environment with other actors, systems, people, and the scary outside world.

Sounds abstract? Until the late 90s artificial intelligence was only done in the lab, now people interact with AI, unconsciously, on a daily basis, for example if they use Google, check their Facebook page or look at banners on the web. Take the company where I work next to my academic job, when I came in 2002, it was a startup of only 15 men with new software and a launching customer [editor’s note: we know that feeling ;)], ten years and two acquisitions later, we have reach more than 1 billion consumers with intelligent, data-driven, scientifically proven, real-time recommendations via digital as well as traditional channels like ATMs, shops and callcenters. No push product offerings anymore, but only ‘next best action’ recommendations that optimize customer value by balancing customer experience and predicted interests and behavior.

What opportunities do you see for startups in the artificial intelligence in this area ?

Well, I see tremendous opportunities, not only for 100% pure AI startups, but for all startups . If you look at the startups in Silicon Valley in high-tech and biotech, artificial intelligence is a major part of the business. Every startup should consider whether data is a key asset or a barrier to entry, and how AI or data mining can be used to convert these data into money. Where I have to note that customers and citizens rightly so, are getting more critical after all NSA issues. Those who can use this technology in a way that it not only benefits companies, but especially customers will be the most successful.

In conclusion , I am curious about how much you you are looking forward to December 13, and what should happen during #projectwaalhalla that would make your wildest dreams come true.

Very much looking forward to it! In terms of wildest dreams: I heard a reunion concert of the Urban Dance Squad is not going to happen, which I understand, but I look forward to exchanging views with startups, freelancers and multinationals on how to create, with the help of raw data diamands and a magical mix of data mining, machine learning, decisioning and evidence-based and real-time marketing.

I will bring some nice metaphorical pictures and leave will double integrals at home.

Do you like what you have read? Get a quarterly update of what I am busy with.

[Editor’s note: An UDS reunion? Sounds like a plan to us]

Originally published at www.sciencerockstars.com on December 6, 2013.