Cooperating with robots
How technology will bring humanity closer together, the growing importance of creativity and how banking will look like in a few decades — these are just some of the questions touched upon by two of the most sophisticated digital banking gurus in the Baltics, Ģirts Bērziņš and Iļja Nogičevs in their fireside chat. Let’s join them.
“I believe the future will come to life much faster than we think now. Because as yet we’re still dependent on the tools that customers have. However, AI will be built into the world outside. So the infrastructure is ready.”
Co-chair of the Finance Latvia Association Digital Transformation Committee, Member of the Board, Head of Strategy, Digital Banking at Swedbank AB
“Yes, it will be nothing like what we have now. Yes, we should definitely think about how it might be, and also think about what would be the baby steps.”
Co-chair of the Finance Latvia Association Digital Transformation Committee, Head of Digital Development at SEB bank
How is artificial intelligence (AI) understood in banks today?
Ģirts Bērziņš It’s a buzzword. Today in banking everything is based on data; it’s a non-material service — money and cash are already a secondary stage. So AI can be applied pretty much everywhere; there are no limits. If I look into the future, I see that there may not be just one robot working with all the customers, but each customer will have their own virtual assistant (VA) or robo-private-banker looking after them, and knowing about their particular spending, and about their lifestyle and habits. So, instead of relying on mass marketing and trying to sell stuff, virtual assistants will make customers want to buy stuff by giving them advice. Here, trust is the most important thing, as customers will have to trust that their VA is working in their best interest. And maybe your personal assistant will tell you that it is in your best interest to go and buy this product from a competitor of your bank. Because for your particular need it’s much better — and since it’s your trusted advisor telling you this, you would do it. And from the open banking perspective, this is the most crucial thing banks need to understand: that you may need to give something up, but you will gain much more back in the shape of trust — the more trust you have, the more business you can do.
Iļja Nogičevs I totally agree with what you just said. For me it’s important to know where I can turn if I have a question to ask. And I feel that this is the status we’ll be fighting for — to be the one you come to with a specific question. And it’s OK if you are not selected for all the services your customers need. I was also thinking about whether AI is just a buzzword or not, and now I actually decided that it’s rather a long-term trend. Talking about it from a more primitive perspective: it’s just robots mimicking humans, and it’s already been there for some years. For example, when a customer sends an insurance agreement including extensive terms and conditions for their insured mortgage, we have a machine to read it and put it into the system. It’s al- ready up and working, and we’re quite good at things like this. Text, voice and speech recognition, this is the basic stuff, but what about understanding unstructured data? For example, we’re getting information from all these questionnaires and putting them into a specific tool and getting results that are already structured, like “your customers are complaining about this, this and that. And if you fix this and that, your results will be much better.” Previously there’d be a few operators trying to do all the data entry and analysis by themselves, which takes ages. That’s my understanding of the AI applications which currently exist. The next stage is machine learning. How do you train an algorithm and how will this algorithm be capable of training itself? And then the last one: the output, and how to bring it to the customer. Either it’s via some sort of hint to the person who is serving the customer, or it’s via a chatbot or VA. And here it’s actually up to the customer to say “no, I’d like to speak to somebody, but I’d like that to be somebody humanish.”
ĢB You are right. I just think the challenge lies in how we think about the development of AI, as we tend to only think about it from the perspective of what we know today.
Making the most of one’s personality
ĢB Our chat is being recorded at the moment, but imagine if it wasn’t just transcribed into text straight away, but that it was also adapted by AI or your VA to fit exactly its reader’s personality. We know that humans differ: some are visual types, and for them maybe it’s better to convert a text into an info-graph or a picture. Others like bullet points and PowerPoint presentations. And some might like formulae! AI will individualize services. By the way, this is how humanity started; there wasn’t any mass production of one-size-fits-all products. And I think of this as maximizing human capacity, in a good way. Not squeezing people out, but boosting their personality. And think wider than banking — your individualized robo-teacher or enhanced doctor with immediate access to a vast amount of data and research. Imagine now that you have a choice between using a regular doctor or an AI-enhanced one. I think that before long you will come to prefer the latter.
IN Again, my thinking is: ‘step by step’. There will be nothing like what we have now tomorrow. Yes, we should definitely think about how it might be, but also think about what the baby steps will be to get there. And maybe I am too cautious, but I would say that, in order for us to succeed, there are a lot of things that need to be prepared or at least done in parallel. I’m thinking here about legislation. And as banks we are regulated quite differently, so we cannot do things like that at the snap of a finger. One can buy a company or services, which then would be integrated into a bank’s offer but then there is the question of the legal burden — who is doing that? Is it you, or is somebody doing it for you?
ĢB I can explain why I believe this shift will come to life much faster than we think it will at the moment. Because in implementing better self-service, we’re still dependent on the tools that customers have. However, AI will be built into the world outside. So, the infrastructure is ready; it’s just a question of use. Self-service will fully disappear after some time, except maybe for those who still would like to be in control.
... and the button which shuts it all down
IN And another thing AI can do: it can make you feel in control.
ĢB That’s a good point, and I have one example in mind. We all know that in the world there are several boxes with special buttons that launch things. I think that in the future there will probably be only one self-service button remaining...
IN ...the one which shuts everything down.
ĢB Yes, the one which stops everything. Because I think that humanity will become extinct the moment it loses control. And this is a real-life example. We know that there
are companies that stopped AI projects because the robots started speaking a language that humans did not understand. And that means you’re out of control. So, I hope that humanity will still manage to keep this ‘stop’ button.
What is the future going to look like?
ĢB I am quite sure that AI and VAs will come very soon, and that customers will adapt to them quickly. And then it will disrupt the whole service industry because you suddenly won’t need to understand exactly how a service provider’s website works in order to use the service, for example. And yes, for services where customers already know what to do, self-service is brilliant. But quite often, even the most technologically advanced customers stop at the point where they need support. And then the gap between the experience of helpline support and an instantly responsive VA is enormous. Self-service has raised the customer’s expectations to a level that they now think everything should be instant. AI will help solve this problem.
IN I agree that this is true with regards to basic products like daily banking. Still, if we are thinking about financing and investment, I have a strong feeling that customers are not ready yet to get this instant advice. This also may be a question of perception — why I am paying so much if this takes so little time? Customers tend to perceive the process of getting advice as more valuable if it is longer. So, I think that maybe it’s better to put it just as assistance, not as a robo-advisor.
ĢB Yes, I was rather talking about customer support. But since you touched the topic of advisory services and sales, I think that selling financing, for example, will stop too. Nowadays payments are already built into real-life situations — that’s the first banking service which is disappearing. Today you will use Taxify or other services—you just use it and your payment is made automatically. Because what the customer needs is not to make a payment, but to get a taxi ride. Also, customers don’t think of their needs as things like ‘a consumer loan’ or ‘a mortgage loan’, but as ‘a house’ or even ‘improving my life’. It’s obvious that with time financing will be built into situations that somebody who arranges the service will take care of. That’s the thing — almost all the services that banks provide today are not primary needs of customers; they are secondary needs. And in the future such services will become invisible for the customer. Except, I’d say, advisory services. Advisory service is a first level need — how should I manage my finances in order to improve my situation?
IN I think this is a question for the customer: “how do you see interactions in your daily life?” Imagine that in the future you come home from your vacation, for example, and your fantastic VA says “I see that you’ve spent less than you thought you would. Would you like to put the money left over into your savings account, so your next vacation will be sooner, or so you can afford a fancier hotel next time?” So, here I see our job as being how to bring in more value.
ĢB Well, maybe not in this stage but the next one, I think that AI will basically disrupt how we understand all areas. Let’s take HR, for example. Today HR strives to maximize value for each employee. In the future, each employee will have a virtual HR assistant, who will take care of their personal skillset. You can’t efficiently develop skills for all employees from a central point. It’s in the hands of each individual employee. And now you’ll have a VA saying “look, there is this vacancy. Knowing your skills and strengths it’s going to be a good opportunity, so now let’s have this module of training...”
IN Or something like: “I feel like you’re lagging behind...”
ĢB Yeah, and “I know where you want to be in your career in five years’ time. So you should be more active in this area. Maybe go to this conference, learn that.” So, I think that this will open up new opportunities, which will maximise the human potential.
IN Because we’re used to thinking that AI will take our jobs, that it’s only about cost reduction and stuff like that. But this is not how it works. The way we’d like to see this is basically ‘how can we help’ — to help the customer or the employee, or the nation itself. And from the perspective of a new customer, what we’re talking about now is in fact already operational in HR — there is technology that can understand your behavior remotely and from all the readings it takes interpret what you like or don’t like, what your preferences are. Imagine the use of this technology from a recruitment perspective.
ĢB I think that the recruitment process will become obsolete too, because if you have your virtual HR assistant in the organisation you’re working in, the assistants from all the different organisations will agree among themselves what the best place for you is.
IN And it could be that your VA will say “Ģirts, now it’s time for you to work in another bank, because it’s a much better workplace.” And our VA will say “do you know about Ģirts? He’s kind of OK — let’s take him.” And today it sounds like a joke. But the question is: will it be a joke in twenty years?
ĢB A lot will happen in twenty years. And here one more thing we should consider in coming back to banking services — banking services are invisibly interconnected. So, if we want to meet our customers, we don’t always have to push our customers to come to us. For most situations, they can be wherever they are — Facebook, Amazon, etc. And banking just happens. Now customers have the feeling that banking happens only when customers interact with the bank (and at the moment this may still be true). But banking is moving away from this. In the past you visited a bank branch something like once a month or once a week, then banking started happening daily via apps. But today banking is not even instant — it’s becoming constant.
Will the Baltics become the next hub for innovation?
ĢB Will the Baltics be the next Silicon Valley? No. It can be a chapter or it can plug into the ecosystem of another region. But it’s just too small. And in terms of becoming a part of something bigger, I think the outlook is rather good. Really finding a niche, like, for example, Tilde have done with small language technologies; I haven’t seen anybody else in the world who has progressed so far with this type of algorithms. But I think any company will start to have a problem if it decides to limit itself to here. They need to become part of some- thing bigger, a part of development of this kind of AI on a larger scale and plug their developments into somebody else’s ecosystem. So if this happens, then we will have a bright future.
IN I think we have a high level of potential, as the customers and the economy at large and the players on the market are all ready to try, test and do this. The Baltics are a good playground (sorry, I don’t want to say Latvia — it’s way too small). But as you said, it’s a very small playground. The only question is about the scalability perspective. It’s OK for trying out and testing, but what’s the next step? If you don’t have this established, you can have a chapter, a hub or whatever, but what’s next? Where will you take what you’ve created and copy it on a different scale? For the bigger players like US and China, it’s kind of easier for them to do that. And here we don’t have a strong partner. Is it Europe? Is it the Nordic countries? Maybe, I don’t know.
The more humanity invests in robots, the more people come closer to each other. Because there will come a time when we don’t need to solve small technical tasks anymore. Then humanity will need to solve creative challenges. And for creativity people will need to move closer. I think that creativity is the next step after Industry 4.0. Yes, today AI can compose music, but it composes it based on the past. It’s too much synthesised. The next frontier will be to disrupt true creativity.
– Ģirts Bērziņš
Customers need to feel like they are the key and that the bank is doing the boring work for them while in many cases staying invisible. However sometimes they would still like human interaction. We can create a super virtual assistant who will know where to invest and what your risk tolerance is, but one still should remember — let’s start from single and simple things and build up trust. And it is our job to make clients’ lives better, to assist them in taking the right decision.
– Iļja Nogičevs
Story published in Magazine Finance in Latvia Nr 1 (September, 2018)