Anything that can be connected, will be connected
2017-10-18 - Guido Jouret, Chief Digital Officer at ABB, speaks about looking for patterns through machine learning and about digitalized products that are not just must-have gadgets
There are lots of terms which describe a digital transformation. Some of them are used interchangeably. Are we still talking about that same process?
In manufacturing, especially in Germany, we are talking about “Industry 4.0”. In the US many companies refer to this as the “Industrial Internet”. In the electricity markets we are talking about the Energy Revolution. I think we are all saying the same thing, which is the impact of digitalization. It is really nothing new. In the late 90s digitalization came to computer companies, to retailers, to banks. At the time it was called “e-business”, not “digitalization” or “Industry 4.0”. There was a strong interest in all of these topics – how to manage digital, how to hire the right people, how to unlock the value.
In a sense the same questions have been around for 20 years now?
At that time “e-business” impacted part of these companies, especially their sales businesses, such as on-line banking. But the current digitalization wave is about a lot more industries and is impacting every part of these industries. In the automobile industry, we are automating the factories with robots and all of these solutions we can provide. The way of producing cars is becoming more efficient and flexible. However, the car itself is also changing. It is becoming a computer on wheels. You see more GPS-self-driving-cars. You have infotainment systems inside the car. It is more an integrated, more digital, more connected type of product. The car is now connected via LTE. It is uploading information every night. Soon we will start to call the car using our mobile phone and there will be no need for drivers. So, the industrial revolution is where we are digitalizing the way we make the products but we are also in many ways digitizing the product itself.
Is it true that Internet of Things isn’t quite Internet of “Things” but Internet of Data?
In many cases the companies are collecting a lot more data that they don’t really know what to do with. But if you are not connected, you can’t get data so people are saying: “Ok, let’s start, let’s just get the data and later, hopefully we will find some good uses of that data”. Ultimately the benefit of data is when we can start making decisions based on it. That is why people are increasingly looking into machine learning and those kind of things, to look through the data and find patterns. The ultimate goal is to start doing predictions, to be able to say: “I know what to do, I am going to make this happen”. As a result you are doing what we call “automation”. You are going from sensing something and analyzing something to acting on it. By closing the loop: sense – analyze – act and back again you will get a real benefit from IoT.
According to studies there will be around 20 billion devices with a unique IP connection by 2021. A widespread digitalization of everything is taking place now?
I do believe it is real. I think anything that can be connected, will be connected. The cost of connectivity is dropping all the time – whether it is the microprocessor or the Wi-Fi chip. Historically, we use to connect the expensive things first. Now we are increasingly connecting almost everything else. If we look at ABB alone we have 70 million connected devices today. We make one million connected drives per year. So many systems we make are connectable by default. I think the number of 20 billion is reachable. It is going to happen because the cost of the connectivity is being driven primary by the mobile phone revolution which creates new solutions and technologies that support this big connectivity process.
Does it mean that the era of “classic” products with analog features is over?
Do we really need these overhead lights now in this room? Maybe not. And if the electricity grid “wants” it we can selectively turn off these things and then we get paid for doing that. So, this demand-response capability is now a feature of the software that is in the circuit breaker. Something can be bought by the customer as a product and those additional features could be unlocked on demand, as a service. It is still the same physical device but some of the value can be provided if the customer wants it. Sometimes the consumer will still buy just a product because she wants just a product and she doesn’t want any of these digital features. Clearly not every light switch will have an internet connection. There is definitely going to be a set of purely analog lights, switches, circuit breakers, motors, drives etc.
Is digitalization in some cases not just a creation of needs and a process to develop new “must have” gadgets?
Yes, sometimes but that’s mostly in the consumer space. For example, I can change the color of my lightbulb automatically. It is nice but it is not really life-changing. I can live without it. But in the industrial space the return of investment, the value digitalization creates – is real. One of the main benefits of digitalized services and products – improved availability, speed, output quality, agility of production, health and safety – are typically the main sources of value. Customers always want that. They always want to improve safety of the factory, cut the cost of producing their goods, get more flexibility and higher productivity. Whether it is done digitally or not. It happens that in today’s world almost all of these benefits are powerfully unlocked with digital technology. If you have a system that tracks where the workers are inside your mine, you are sure that you didn’t forget some people down in the mine and that you provide air down there. But it also saves energy because if you know that where there are no people in the mine you don’t have to supply any air to that part. We are talking about really important things here. In most cases products that come from digitalization are not just nice-to-have gadgets.
You have mentioned the value digitalization creates. What kind of benefits does ABB Ability unlock?
ABB Ability is the name we give to the package of our digital solutions on our platform. What is a digital solution? It is hardware or software, sold as a product but sometimes as a service. And we call these “solutions” because they are designed to solve a customer’s problem. “I want to schedule the production of things in my factory. I want to monitor my ships when they are out at sea.” Each one of those is an ABB Ability solution. We have more than 180 of such ABB Ability solutions today and we can say with confidence: “Dear customer. If you want to improve the energy efficiency in your mine, we have the solution for that. If you want to run an electricity market, if you need to make sure you provide reliable power in the future, if you need to run a digital trading market, we have a solution that does that. If you want to coordinate distributed energy resources like diesel generators, solar panels, winter installations and make them all look like the one power plant. We have software to do that.” So we have these 180-plus solutions and some of them are for utility customers, some for manufactures, some of them are for the oil & gas sector.
ABB is cooperating closely with Microsoft and IBM. What is the framework of this cooperation?
We have decided that all of the new ABB Ability solutions will use Microsoft Azure technologies. What is exiting about Microsoft is that they are building technology that will let us run applications not only in the cloud but also inside the factory. That is quite interesting, especially in utilities where our customers don’t typically want data to go to the cloud for security reasons. We are using IBM’s technology in a way we call “intercloud”. Customers often have asked us to interconnect various systems because they have ERP, they have something else from other companies. Every company always has existing applications and systems. And some of the providers will say “it is not a problem, just move your application to my cloud”. But in the world of cloud, what does that really mean? We have a different answer. We say: “Actually, your application can stay where it is, let’s focus on connecting these applications, lets create interfaces so they can communicate and share data.”
Ulrich Spiesshofer, ABB CEO (left) and Guido Jouret. Fot. ABB
And that is what ABB is doing with IBM Watson?
Yes, for weather forecasting we send information to Watson which then helps produce recommendations for what is the best time to do maintenance on a wind turbine. You want to do that when you know the wind will not be strong. The other idea would be to use the data for what we call quality inspection in a manufacturing plant. We send photographs that are taken by a robot to the IBM Watson system, which looks for defects and tells us where the problems are. It has learned how to spot some of the manufacturing defects on an assembly line. We are not moving IBM Watson technology into our data center nor are we moving our robot automation system into an IBM data center. We are just interconnecting the clouds. IBM also helps us with developing specific new solutions for vertical markets. IBM is historically very strong in certain vertical industries and makes software that in many cases is very complementary to ours. IBM is very strong in IT, ABB is much stronger in what we call OT or “operational technologies.” Customer want to connect these two worlds. They want to have the automation system from ABB talk to the finance system or some kind of ERP-type system. Companies like IBM can help to create that connection for us.
You have noted some concerns from utilities about security aspects. Are you not afraid that when everything is connected, someone will find some kind of skeleton-key and crack many of the systems? How does ABB protect devices, especially connected devices?
We have been using the latest and greatest cybersecurity technology already and we will continue to add to that. The latest additions would be technologies based on machine learning to detect new kinds of threats because quite often the problem is that you are only able to stop things you have already seen before. But you don’t know what else might be coming.
Artificial intelligence can solve that problem?
Yes, the technology that is based on AI learns to recognize the pattern of attacks, looking for the common patterns of specific threats. These technologies we are sourcing and putting into our devices for example. However, the customers actually have two questions. First on cybersecurity and we are telling them that these kind of technologies will defend the product. We encrypt the data as it is being transmitted and we encrypt the data when it is stored. The customer holds the keys. I think that is quite well accepted. But sometimes customers have questions about data ownership, which is different. In the consumer space it is called privacy. In the industrial world we typically refer to it as data ownership–who owns the data. Meaning if we monitor all these robots we can learn a lot about how the customers produces their products. That can be a problem for some of our customers.
You might learn a lot about what I’m doing and that could be a risk for me?
Or the customer would say that if ABB or any other provider learns how to improve their production, will this provider then sell the same capability to competitors? In other words: you have learned from me. Are you going to let my competitors benefit from all the hard work that I have put into this? The cloud business has not been very good at explaining to customers what their rights are, what data are we are gathering, what we do with the data, and (if you choose no longer to be a customer) what happens to the data. That is why ABB wrote what we call a “Data Manifesto.” It’s not a law because at the moment there are very few laws about ownership of data for industrial companies but it is more like a set of best practices. What we propose is that companies such as ABB, which are providing industrial IoT solutions, would adopt these principles. They are inspired in large part by some of the principles or regulations that apply to healthcare. For example, we believe it is important to separate your identity data from your measurement data–who you are versus the data coming from your robots for example. In healthcare, typically a doctor who performs clinical trials or studies only sees the patient’s number, she doesn’t know who that patient is because that information is in a separate system. By separating the identity and measurement data you reduce the risk. Very few people should need to access the identity data but larger populations may have access to measurement data. We believe that there are three kinds of data: identity data, measurement data and then the learnings–what we’ve learned from looking at all of this data. We wrote in our Data Manifesto that we will protect the customer’s data at every level with encryption, that the customer has only access to their data, and that we will delete the customer data if the customer asks for it. This includes both identity data and measurement data.
And what about the learnings?
We can’t unlearn what we have learned. If we have made our robots smarter thanks to working with the customer X. We can’t make them less smart again. Many of our customers are ok with that. If we make our robots better by working with company X then customer Y who buys this robot, and who is the X’s competitor, would benefit from that learning. I have personally asked a number of our customers how do they feel about that. And they say: “why would I want to be handicapped by buying robots which are not as smart as they can be? I wouldn’t want that. And I understand that you are learning from me but you are also learning from them so we are also learning from each other.” The value goes both ways. That is true for the software in general, even for software on your laptop. The input coming from many users of that software is making the software better for you. The automation world is no different. In digital world that’s same. But we believe the opportunity of talking about the data ownership is something that we need to do more of. We think that it is important to increase the adoption of cloud. Many people are holding back not because of the concerns about cyber security but because of concerns about data ownership. Because it is not clear.
What kind of concerns do customers have about digitalization? What are the barriers for them to introduce new technologies?
I think concerns would be around whether a solution really delivers the value we claim it will deliver. The barriers would be around expertise and access to the people who have that expertise and are able to deploy digitalization.
You talk a lot with industrial companies. Were there many situations when you have heard: “I don’t care about digital, I don’t know what you are talking about”?
Speaking about a wide range of industries, I am pleasantly surprised. In many cases – in Oil & Gas, Automotive, Food and Beverage and others–I say: “Would you be interested in having a conversation around digital?”, I’ve had an almost 100-percent success rate in getting that kind of a meeting (laughs). I think it is because we can experience digital all around us. If you drive an electric car you can see this evolution, if you look at someone’s home automation technology, if you are looking at what is happening with robots in general, you would think that things are really starting to change. I believe it is a combination of greed and fear. Greed which is: “What are the opportunities for me if I pursue digitalization?” And fear which is: “What will happen to me if I don’t”? And there is of course constant competitive pressure when you realize that standing still and not doing anything is not going to save you. You have to find some way to keep renewing your differentiation.
Poland is well-known for its IT-specialists and software community. Does ABB’s footprint in Poland confirm this?
Yes, absolutely, we have a total of 2,500 employees here in Krakow, a large part of which are working on software development. We have great universities nearby, which create talent we can attract and hire. ABB has its Polish Software Development Center located over here, its Corporate Research Center as well as IT teams, which are a part of the global business services organization. And we shouldn’t forget about IT specialists who work at the factories in different ABB locations across Poland.
ABB in Poland celebrates its 25th anniversary this year. What is the contribution of the Polish organization to the ABB’s global success?
We have a large production footprint over here so a lot of devices that ABB sells around the globe are made in Poland. But the Polish organization contributes to our business services, software development and R&D as well. In our Corporate Research Center in Krakow researchers are experimenting with – among others – the next generation of wireless communication technologies, software algorithms and analytics. They are also working on mobile application development. In particular we do some of our best new cloud solutions over here, for example Asset Health Center, which has been developed by the ABB’s team based in Krakow.