Interview with Maciej Stanusch.
Maciej Stanusch has twice been listed as the most creative entrepreneurs in Poland. In 2007, he received an award of the association Silban for an innovative business idea and, in 2010, the award Regional Leader of Innovation 2010. Twice nominated for the Ernst & Young Entrepreneur of the Year award and for the award Innowatory Wprost 2014. Associated with the IT industry for 30 years. A graduate of the University of Economics in Katowice and many additional courses and trainings. For many years Maciej Stanusch was associated with the industry of CRM solutions. He is the author of several publications related to CRM, he created the first Polish vortal dedicated to CRM and is the co-author of the book “CRM – Przewodnik dla wdrażających”. He designed CRM systems for Polish manufacturers of the software: BPSC, Teta and Clix Software. He has participated in CRM projects for companies like Compaq Computer, Microsoft, SAP, Schneider Electric, Rockwool, Sodexo, Acer. Since 2003, he has been dealing with the subject of artificial intelligence (AI).
Glivia Team: Let us begin with what is artificial intelligence?
Maciej Stanusch: This is a difficult topic because it is sometimes unclear what intelligence in general, so-called natural intelligence, is. There are many different approaches to artificial intelligence and many definitions. The most common definition of artificial intelligence is software that has functions previously reserved only for humans, functions that cannot be easily algorithmized, such as text understanding, face recognition, decision making or solving problems such as proving mathematical theorems.
GT: How can artificial intelligence be brought into marketing and management?
MS: There are a few areas where applying artificial intelligence to marketing can be very useful. Let me start with a very mundane thing related to direct marketing. It is very common for a company that does emailing campaigns to have many different databases. For example – we have our database and we buy a new one. It would be good to send information to potential customers from the new database, except for those to whom we have already sent the offer before (and some items from these two databases may be duplicated). This is where the problem arises, because usually databases have information stored in different formats. Someone may enter a company as: ABC, someone else may add ABC LTD., or ABC Limited. It may be the same company, but it will be treated as three different entries. Manually finding duplicates is not a big problem when we have a few hundred entries. But if we already have several thousand of them, it is a gigantic problem.
Without the right software that is equipped with intelligent recognition of these duplicates, that can understand that a company may have different names, the whole operation is extremely difficult. This is the first solution that uses artificial intelligence – cleaning databases and finding duplicates in direct marketing.
The second application is customer behavior analysis. For example – we have a bank where customers use different types of services. Depending on what their material status is and how old they are, they use a different package of services. There is software that can analyze customer behavior, for example, a student opens his first account, then goes to work, so he will probably want to have a credit card or maybe some deposit. Then he will take out a loan for a car and for an apartment or house. The system is able to predict at what point the customer will look for certain products. Having such software it is possible, in a way, to anticipate the customer’s decisions. Thanks to this we can also get ahead of the competition.
GT: Thanks to the automatic analysis of customers, we save a huge amount of human labor that would be needed to perform such a task.
MS: That is one thing. What is more, we can, as I have already mentioned, anticipate the customer’s decisions. In some banks (unfortunately, not in all of them) it happens that a client who has a current account at some point “somehow automatically” receives a credit line in the account – the system has determined that the client will probably want to use such a service (and has calculated the risk by the way), and the client only has to accept this proposal. The bank does not ask for any documents, it is only the system that determines as a result of operations on the bank account that another service can be sold to the customer.
GT: This reaches into relationship marketing in a broad sense.
MS: Absolutely. Except that everything is based on a computer system that constantly analyzes the behavior of our customers.
GT: A rather spectacular application of artificial intelligence are the so-called chatbots. Can they be used in some way in marketing?
MS: Yes. Let us start with the fact that it is becoming very popular to automate customer interactions through software that can communicate in natural language. One of the problems that such software solves is the automatic response to customer emails. If per day we receive dozens or even hundreds of emails it is not a problem, but if they come a few thousand, it begins to be an enormous problem. Software was created, which can automatically analyze the content of such email and automatically respond to it. Someone, for example, writes an email to the company about the price of a product, and the system, based on its content, automatically responds to it. This means great savings of time and work for people who deal with customer service. What’s more, we can automatically save it in our CRM system and pass it to the right person in the company, who can start the sales process.
An extension of such functionality are programs for conversation between the user and the computer system, the so-called chatbots. Thanks to them we can “talk” (or rather chat) with the machine. The user can have a conversation with the computer system without knowing about it (if it takes place via a website). These systems are not yet super-perfect, but within a certain domain – for example related to a product – they work very well.
The next stage is adding speech recognition systems to conversational programs and replacing traditional call-centers, with virtual ones. If someone calls them, they will talk to a machine. This is the future, of course, but speech recognition techniques are getting better and better, and it is a matter of years at most when this type of solution will become widespread.
GT: With that said, the problems and concerns can be multiplied in this case. Customers will ask questions very imprecisely, sometimes they will react emotionally. Will machines really replace live consultants?
MS: It depends on how extensive the knowledge base that such a system uses is. If it is small, it is indeed very quick to discover that it is not a real person. However, there are many projects, run by my company among others, whose goal is to create a gigantic common-sense knowledge base. It consists of hundreds of thousands of different pieces of information, thanks to which the chatbot has knowledge about the world around it, it knows what is going on in politics, what the weather is today, what is on TV, what new movies are in the cinemas. This way it becomes a substitute for real intelligence.
What is more, the latest solutions of this type are able to combine unrelated facts by means of deduction. This is already a substitute for the real thinking we have to deal with in the human brain. This is just the beginning, but these programs are becoming more and more perfect.
In the few areas I have mentioned, the use of artificial intelligence will grow the fastest. Personally, I think that in a dozen years or so such an area may be robotics, but it is difficult to predict the future.
GT: What can robotics be used for?
MS: In general – for customer service. It is not only about software that communicates with the customer via a website or a phone line, but also about devices that deliver goods, maybe even serve the customer in the store. Vending machines are a certain substitute. The future will be communication with such a device through speech, not just a keyboard. Who knows, maybe one day we will have robot door-to-door salesmen? But for now, it is science fiction.
GT: This leads to an important question – what is the market potential for AI-based products?
MS: I will honestly say that while the products themselves are getting better, there is a lot of resistance to innovation in many countries. However, it seems to me that we will not escape this. It is very difficult, for example, to find employees for a call-center for little money. Therefore, at least because of cost reduction, such solutions will be adopted. Maybe instead of three real salesmen it is better to employ one virtual one that will partly facilitate customer service.
GT: Thank you for the interview.