Reinventing the Digital Workplace at Swiss Re

Michael Krigsman: Welcome to Episode #214 of CxOTalk. I’m Michael Krigsman. CxOTalk brings together really the most interesting, innovative people in the world for in-depth and genuinely substantive and meaningful conversation. I want to thank Harmon.ie for sponsoring today’s show. Harmon.ie has a set of products that bring information together from multiple sources and make it easy to consume on the desktop. And, I also want to thank Livestream, as always, for being a great supporter of CxOTalk.

We have a tweet chat going on right now using the hashtag #cxotalk, and I hope you'll join in, and you can ask questions of our guest during this live show. And we have the pleasure and the honor today of speaking with Rainer Baumann, who is the Chief Information Officer for Information at a hundred-and-fifty-year-old major, large Swiss insurance company called Swiss Re. Rainer, how are you?

Rainer Baumann: Good! And you?

Michael Krigsman: I am doing great! Thanks so much for joining us today! Please, tell us about Swiss Re.

Rainer Baumann: So, Swiss Re is one of the world’s largest reinsurance companies. Reinsurance might not be familiar to many of you. However, typically you work with your insurance companies, which give you coverages for your home, for your cars. However, if they are working in some countries, or focusing on specific areas, they typically need to balance the respective risks. They reach out to companies like us, where we help them for extreme events to cover for that.

Michael Krigsman: And, what are your role and your mandate at Swiss Re?

Rainer Baumann: My team and I are taking care of all information-related aspects of Swiss Re, such as collecting, storing, and managing all information. And we are focusing on exploiting and leveraging the latest possibilities around digital analytics to get most out of it for our business.

Michael Krigsman: You talk about the company as being a knowledge company, and knowledge is central to the type of insurance business that you’re in. So, can you elaborate on that, and tell us more about what reinsurance is, and how it connects to data?

Rainer Baumann: Reinsurance itself is not [so] much about interacting with policyholders ─ that's what we call individuals like you, Michael, who are typically purchasing insurance, or with SMEs who are also buying insurance. We are focusing more on supporting the large and smaller insurance companies on how to deal with the customers. So, we give them an understanding about the different risks: how they calculate it and how they work with it. To give you a very simple example: If you want to purchase life insurance, and if you're a very healthy person, that's a very straightforward thing. And most companies will just give you a nice offer. However, if you have a few medical constraints, it might suddenly become very complex. How do you price this risk? And "the risk" means the price you actually need to pay for this insurance. We help the primary insurers identify that.

But, there is also much more than life insurance. If you go, for example, to industrial insurances. When you have very complex facilities, we do something we call "risk engineering." You go out to these facilities, [and] try to find out: what can go wrong; what the impact might be. And this needs tons of experts.

So, if you look into Swiss Re, you'll only find a few guys focusing on bringing insurance, or reinsurance to our customers, but tons of experts from many, many different areas. Tons of engineers, physicists, [and] mathematicians who build the brightest models about earthquakes, weather, [and] people aging. All these graphs [word?] make us into a melting pot of experts, who give us a nice and interesting cultural environment.

Michael Krigsman: So, your business is not focused on end-consumers. You are supplying a backstop primarily for other insurance companies, and much of what you do relates to gathering data about different aspects of the world, integrating that data, and finding ways to manage that data so that your in-house experts can make the most informed choices.

Rainer Baumann: I would add formally that we try to understand the world. Sometimes, there are great models that you can apply to understand the world. Sometimes, you need tons of data. And sometimes, it is a lot of long-term expertise, learning, and the latest research around how, for example, medical costs are creating one of the biggest unknown problems: Will they increase by 5 or 10% annually? Or, will they go down? And if you knew the extrapolation of the treatment costs over the next decade, it would be a very relevant factor for understanding how to price these actual health costs. 

Michael Krigsman: So, you are the CIO for Information, which I’m assuming means that this information challenge is an important part of your mandate. And so, how do you think about bringing this information in, and [about] the nature of the workplace and how your workplace is evolving to make it easier for this group of data experts to carry out their job?

Rainer Baumann: First of all, luckily, my folks don't need to be the biggest experts in these different insurance disciplines and the underlying knowledge areas. When we talk about bringing in data, we are not only the experts in how to work with that data and on methodologies but also leveraging the latest technologies. When it gets to mathematical models, for example, or how you calculate weather conditions, then we have our engineer-physicists who know this much better than we do. However, when you have …

Let me give you a little example. When you get a claim after something happens, then classically, you will have somebody typing this from an email into a system. From that system, someone would review it, and put some more information into another system. However, everything you get, you could extract more or less automatically, and digitize significant part of the processes. But, you need to leverage quite a lot the capabilities that are slowly emerging in the field.

And it is the same as [for] understanding different risks. There is so much information there and we only use a minuscule portion. But, the majority of that information is particularly unstructured ─ is unrelated, meaning it comes from very different sources. You need to put it into context; understand it. And no, we do not employ hundreds and thousands of people. You need to do this in an intelligent way. And, and that’s where our focus is.

Michael Krigsman: … Which, of course, then raises the question of how do you do this?

Rainer Baumann: [Laughter]

Michael Krigsman: [Laughter] And that’s the magic question.

Rainer Baumann: [Laughter] And, two days later, we probably wouldn't have found that silver bullet yet. I would say we apply all the best practices around. When you, for example, acquire external information, you first need to make sure that you understand where it's coming from and the context, and enrich [?] history. With respect to master data, we put it into a context where we have structured information, try to have the right models, and that where you have unstructured information, get them in a way you can work with them.

But, all of what I just described is very much technical. The challenge ─ yes, it is technical. But, the bigger challenge is how do you create actual business value out of it, and how do you enable our how we call it business ─ the experts in the risk functions, underwriting, claims functions ─ to take true business advantages out of this information.

Michael Krigsman: So, you get in this mass of unstructured data, and maybe you can tell us a little bit more about what that kind of data is. But even more importantly, then, how do you make use of it in order to add the business value, and make it useful and in a practical way? I mean, this seems to me like a … For your company, it seems like it must be a strategic competence to be able to do that very well.

Rainer Baumann: I would even go a step further. It is the strategic competence to understand information and leverage it. Or, to do a little historical exploration: How did insurance, over the last hundreds of years, define the risks and estimate the right prices? You looked at what is called "actuarial models," which is nothing else than the development of the respective claims, a bit simplified. So, you looked, and you had a particular risk: How many losses did you have with that risk. And how did this develop? And with this, you naturally assume that the risk develops somehow similarly. Or you add a few parameters and say, "Oh, there are now more interactions so it will have a slightly higher claims rate or a higher cost per claim." But, these developments are naturally extrapolations; they do [not] respect the individual risk. So, we say, "You move from, let's say a statistical risk, to a more individual risk. And you move from a historical perspective on the risk to [a] more forward looking – what most people would say predictive, or a predictive model.”

There are also sometimes limitations to how predictive you can be, depending on the information you have at hand.

Michael Krigsman: You talk about smart analytics, and what is that? I mean, I’m so interested in this issue of how you manage the flow of data, and how you organize yourself around that as well.

Rainer Baumann: Smart analytics, advanced analytics, digital analytics ─ whatever magic word you want to use, [they] are just interchangeable words to me – is leveraging intelligent algorithms paired with massive computational power. By the way, most of these algorithms have been here for ages. So, if you studied computer science in the nineties, you have heard most of it. But, most of what the people said is not usable in reality. So, with sheer computational power, we can now do many interesting things; and this starts by understanding information.

A very simple example is when you have a sentence from a person. Is that person angry or not, which is a sentiment analysis. You now could try to analyze the grammar, and if one word comes after another. Or you add a hundred thousand or millions of examples, and compare them, and say, “Typically, this is angry. Typically, this is happy.” And you can apply different kinds of these algorithms; and you find out, "Oh, this is a happy customer or an angry customer." And, this sort of sentiment analysis is just a tiny glimpse.

Now, think about when you have a text, and in your text, you have written something about … Let's go into a property ─ a facility of a big manufacturer. When they write in the assessment, "We had a local fire brigade for over ten years, but didn't have a good experience. This is why we fully trust in the fire brigade from our town.” So, we immediately know that they don’t have a fire brigade, and we hope the town is very close. Otherwise, the facility will burn down when they have a large fire. For a computer to understand such a sentence, and really get, “Oh! So they do not have one.” But they said, ‘We had one over ten years [ago].' There is not even a negation in the whole sentence. All this is tough to understand.

And besides this natural language processing part, there are also tons of other things. For example, when you want to understand forms, you scan a form. It is tough to comprehend. Where have people done across; where have people done, or how have they filled certain parts of it. Now, funnily enough, we can apply algorithms from genetic engineering that were developed from DNA analysis to detect patterns in forms. These are graphical algorithms. So, you do not apply them to text; you apply them to graphics. Based on these patterns, you can pre-process the information in a way so that afterward, you can go much further.

But, it also goes towards classical quantitative methods enriched with newer capabilities. Artificial intelligence can also be applied to well-structured data, to find hidden insights which you wouldn’t have found yourself.

Michael Krigsman: [...] There’s also a human dimension to this, which is: You come from a very well-established industry, the insurance industry, in a company that’s 150 or more years old, and so, how do you ensure adoption? How do you get your workforce to change the way it works, in order to absorb these new ways of thinking about data, analyzing data, incorporating data, or making decisions with data?

Rainer Baumann: And this is a huge challenge you're bringing up here because it's not just about analyzing data. If you think [about it], we typically have experts whom we have had with us for years. So, at Swiss Re, you can find a lot of very experienced intel people. They are confronted with an industry that is suddenly changing because it’s getting more products in there ─ for example, new cyber defense products. Then, they get a work environment that changes. Ten years ago, they never used any sort of iPads to go to customers, write in notes, chat and interact in high frequency with their customer [when] developing a contract. Then, you suddenly have new models of transparency, and understanding of how risks are developing and where it’s going. And then you have all these new possibilities with data. So, all of these things suddenly come together.

And, in this environment, to be fully transparent, we don’t have a gold nugget. And we heavily struggled with [on-boarding] all of our workforces onto the journey by adapting different capabilities offered. And, the solution and path we are taking at the moment, is to provide different possibilities and opportunities to them and let them choose, to some extent. [We] tried sometimes to impose a certain way of working. We failed pretty miserably.

If you think that you understand how the folks in the company should work, you are either working in a factory, or you might have to rethink your position. And that's what we now do constantly: Try to provide opportunities; listen to them; bring the whole "consumerization" into the game, meaning, like, enabling them with their own phone tablets at work. We also have a big mantra, which is, "Own the way you work." So, people can work at home, while traveling, and with that, have a much more attractive work environment for them. At least, we have had some pretty good feedback on how we are exposing these capabilities to our employees now.

Michael Krigsman: So in this case, when you talk about the digital workplace or the future of work, it mainly comes down to convincing these experts to adopt new ways of working that enable them to take advantage of the data - the unstructured data - especially that which you're able to supply through various types of data feeds and analytic tools.

Rainer Baumann: Yes, that’s one part of it… You can also see this classical adoption problem with the latest technologies when you… and let me perhaps switch to another industry ─ if you go to lawyers. Lawyers, especially in their early years, spend tons of time finding the right information somewhere, preparing it, and distilling insights out of court case decisions. There is a small portion that is truly high in value-adding: sitting with the customers, listening to them, understanding where they are, defining a strategy on how to tackle the legal challenge. And, that part of thinking, this value-adding, probably remains.

But, the other part that I described initially. You can substitute more and more [with] technology, leaving you on one side with a problem that the other people don’t have the possibility to learn and get into it. And for the well-established lawyers, let's think like this: You for sure have a lot of investment into preparing this show. If you needed to run ten such shows every day and having five minutes to prepare it, I’m very sure your life would be pretty stressful.

Michael Krigsman: Yes. So, this type of change that you're talking about must feed into the ultimate thinking of the business model, right? So, maybe speak to us about your thinking [on] the transformation of your business model, and then this use of technology, and the adoption of the different ways of thinking, and the various ways of working; how does that feed into your business model and your broader, longer-term strategy?

Rainer Baumann: Now, these are very broad questions. Perhaps, I may need a few attempts to answer it nicely. First, I always tend to say, “The reinsurance industry is probably the last one to get hit by trends.” First, you see it somewhere in commercial retail, then very, very late in retail banking, then investment banking, then retail insurance. The last is reinsurance. Why is it like this? The answer is because we are a business with very infrequent interaction. Typically, the interactions are semi-pleasant. At least, I don't remember people saying, "Oh, cool! I crashed my car! I can talk to my insurance!" And also not, “Cool! I need car insurance!” and “Cool! I may pay a thousand bucks a year for that insurance! I’m so glad to do that!”

Michael Krigsman: [Laughter] I once said to my accountant, “Doesn’t he ever call with good news?”

Rainer Baumann: [Laughter]

Michael Krigsman: It's the same kind of thing.
 

Rainer Baumann: But, this is … We are in a, let’s say, semi-positive, low-frequency-interaction business. And, having said that, if you are then even second in line, or supporting the companies in such a challenging environment, it’s even more extreme. So, our business model probably evolves much slower than most people would expect when they look at the other end. However, it becomes much more transformed through what is around us. And, “what is around us” means the understanding of the world and how the word is changing. If you look, for example, at the kind of risks we have today, they are so different than the risks we had twenty years ago.

If you think, for example, about cyber risks. There is sophisticated risk that the global economy goes down for several days because of a cryptographic breach somewhere in cryptographic functions. No one knew about this weakness, and it could have been exploited. And then our economies are down.

If we go [back] into the early nineties, what could have happened? You could have had a local power outage. You could have had some instruments that were not produced well, so that [in] several areas, or several places, you have issues. But, this connectivity of the world changes the whole thing. So we suddenly need to understand that the risk is much broader, much more interconnected. And, this also asks us to think, "Can we offer insurance, for example, that is more tailored, more short-term, so that you don't get very long coverage? Could we also partner more with our primary insurers, and the bigger customers, and also individual customers so that they prevent things more?”

By the way, this is also a great chance for us. The digitization shortens the distance from a reinsurance to the ultimate customer significantly. Suddenly, we have a chance to get in touch with them; get real handy signals from them; and also shape their behavior ─ influence them. Because risk management [and] risk reductions is a lot about influencing and assuring that things do not happen that we don't want to happen. And, as an insurer, we are most happy if we can make sure that nothing goes wrong and nobody has a bad experience.

Michael Krigsman: So, when you say that digitization shortens the path or the link to the end-customer, maybe elaborate on that; because again, that is another aspect of your business, which is entirely different from the past, where you didn’t touch those end-customers individually.

Rainer Baumann: Now, let’s start, perhaps, from two angles. The first is how we can influence claims, and how we can leverage more of the potential to get close to the customer. On the claims side, it's a lot about prevention. In the past, we were highly dependent upon insurance companies, or others to help with prevention. Today, we have some insights, whether from the data we can get from our primary insurers or other areas, that something’s not going the way it should be.

So, let's think again about industrial situations. So, you have a company employing 10,000 workers on chemical plants, and we find out that they are running their chemical processes at very much at the edge, which is pretty critical not only towards risking the facility, but also the employees there. When we see that, we can suddenly reach out to them. In the past, we just had the statistics: "new;" "old." Typically, if you produce these types of chemicals, you'd have on average an explosion every thousand years. And if the facility is that big with that blow-up rate, this is the premium. Today, we see, "Oh! The operator on that site is running the plant pretty riskily. He needs to cool it down more,” or whatever. And, what I'm talking to you about now is also very much the future. Today, we don't have that much insight ─ live IoT data, but we’re getting closer.

And, I promised you before that I’d also look at it from a second angle: How we can address the individuals like you and me? Interestingly enough, only about a third of the insurable risks in the world, which would be natural to insure, are insured. So, I am not talking about the coverage for mobile phone glass breakage. I'm actually not sure how much sense it makes to insure your mobile phones from breaking. At least, the insurance industry is thankful for the revenue you're creating, but for you, this is probably not worth insuring. However, there are many things: if you get disabled; if you die; or get seriously ill; or if your house burns down; if there's an accident; or if you get into a legal dispute.  Here, we see a high level of underinsurance. As such, we try to make the people aware of this underinsurance because this is exposing them significantly to risk.  You, living in the US, know well how the US health care system today, or even in the future, with the ongoing changes, will take care of individuals, and how you need to take care of your pensions. Many states have different regulations. But, this is your business, and if you don’t take care of it, then you might have significant challenges in the future.

Michael Krigsman: I want to remind everybody that you’re watching CxOTalk, and we’re speaking right now with Rainer Baumann, who is the CIO for Information at the large, Swiss insurance company called Swiss Re. And, I want to especially thank Harmon.ie for sponsoring this episode. You can send questions to Rainer using the hashtag #cxotalk on Twitter right now.

Rainer, so we haven’t spoken directly about the topic of innovation. And, for a company that is 150 years old, and that is in an industry that’s changing so rapidly right now, I have to assume that innovation is very top of your mind. Can you talk about innovation, which I assume also is the kind of glue that holds these various pieces together that you’ve been describing?

Rainer Baumann: Innovation, in fact, is one of our core paths in the company. Innovation, for us, doesn’t come from tough, stringent, managed research projects. Innovation comes from the entire company. We aim at creating a lot of freedom, and we invite people to think about different opportunities. So, collaboration, open problem-sharing discussions are at the core. You could even say that certain kinds of edgier working environments, which some tech companies have are not new to us since we try to foster those kinds of exchanges. And, when we then look into the individual activities and how they worked, and very often during a lunch talk, or during a discussion with a primary insurer, we come to the idea of, “Oh, we could, or should; and try out something.” And, with our expertise, we are typically able to grow these things.

Naturally, we also have the typical top-down challenges that you, for example, see, "Oh. The population is aging more and more. How do you understand this?" And, because the aging of the population is a huge financial risk to life insurance, [and] pension funds, you need to be somehow able to calculate it. And then, you suddenly find yourself in more tailored discussions. And here, I think we can be fully transparent. Whenever you have a very tailored focus, innovation becomes much more difficult than when you let it come through the organization.

And similar to what I referred to before, tech companies. Most of our biggest innovation came from smart people in the ranks figuring out something with their passion.

Michael Krigsman: So again, that’s a cultural dimension. How are you thinking about the culture, and changing the culture of this 150-year-old company to support this model of collaborative and innovative thinking?

Rainer Baumann: So, on one side, we are lucky that we already have such a collaborative environment. On the other hand, we also have these generation changes, where some are coming in and wanting us to work in new ways. We have tons of digital tools supporting them. However, they don't fully meet the needs of the more experienced generations. So, we try to bring them together with most of the best practices we find out there, whether it is community events ─ physical [events], not like a CxOTalk, but in the Beta science community, for example, around smart analytics. Every second week, we have a talk about a certain topic. In four weeks, we’re about to have a huge conference in our main conference facility, where we have several hundred people getting together and thinking about different problems. We sometimes have competitions, or we conduct hackathons. And, mixing these different elements allows us to embark a bit onto this.

However, we also face a huge challenge here. Since we have this more brain-heavy work, there is a part of Swiss Re that is highly transactional. If you look into our service centers, there are hundreds of people doing operational processing. And we try as much as possible to pick their brains, not just leaving them to dummy information work. We try to do this by engaging them more locally into discussions and opening up the topics. In just two weeks, I’m going to one of our service centers, and have large, group discussions [that] last the whole afternoon where everybody might come and participate, and discuss how we can evolve our processes, products, whatever.

Michael Krigsman: So you’re basically trying to engage people at every level of the organization, essentially.

Rainer Baumann: Absolutely. If you look to our Group CEO, Christian Mumenthaler, who is one of the youngest group CEOs in the industry … I hope he’s okay with me saying that. He’s really interested in tech and he always wants to have the latest features. [He] naturally breaks down these barriers, and enables us to talk very directly on all levels. And, especially since we [have the attitude] that everyone in this company can make a difference, we are so open to doing that.

Michael Krigsman: We have only a short time left, and there are a few things that I still want to talk with you about. But, very briefly, can you share your thoughts on some of the next generation of technologies that relate to insurance, such as mobility, the blockchain, things like that?

Rainer Baumann: For sure, Michael. First, most technologies will transform the way we look at the world. Or, in other words, the change of the world itself is the transforming factor for us. With IoT, we will get more and more insights about how things are working. With this, we understand risks much better and can influence them much more closely. Mobility, or what I would call "connectiveness," allows us to stay in touch with whatever is out there. Here, we talk very often also about connectiveness of people, because IoT is more focusing on devices, on industrial facilities. Sometimes also people, but people, they have much more than their fitness tracker and their intelligent clothes. They're very social; they interact; they leave tons of tracks everywhere.

By the way, just to give you one very nice example: There was an insurance company in the UK that shifted their model for estimating your driver risk entirely just by analyzing your social media footprint. And if you were well-behaving, they assume you're also a good driver. If you have been a "special" character in social media, they assume this special character will also be replicated on the street, and priced accordingly. This level of involvement is not especially appreciated by the regulators, but those are the kind of technology changes [I'm talking about].

We also see a lot of financial technologies from the fintech area, which influences the asset management paying part, like blockchain. And, what other technology have you been thinking about, that would be interesting for you?

Michael Krigsman: Well, maybe digital payments. That seems like it might be relevant.

Rainer Baumann: Digital payments are especially interesting, from my perspective, in two areas. One is that when you get more into insurance-as-a-service on-the-go ─ so pay-as-you-drive, pay-as-you-live─ then you need to have very small payments to capture these respective choices. So, [through] electronic ways, new digital payments enable that.

The other side, micro-payments ─ small payment possibilities ─ are also the key to opening up insurance solutions to areas of the world which have very limited access to insurance. So, if you think about Africa, how can you provide microinsurance to farmers? We’d offer respective payment solutions, and besides this opportunities, this is always, "Sorry, we are an insurer. We are always thinking about the threat I mentioned." It's a huge threat, because if something goes wrong with digital payments, you have a scalability factor. It's much harder to copy notes from a national bank. Whereas doing something nasty with digital currency, we can all go and be very creative about that.

Michael Krigsman: Yeah. So, I can see the concern. We have just about five minutes left, and I know that you’re working with Harmon.ie, and Harmon.ie sponsored this episode. CXOTalk doesn't have commercials, and so I’m always very grateful to the people who support it. So, tell us about what you’re doing with Harmon.ie?

Rainer Baumann: So this is very funny, by the way, because today, we had Harmon.ie exactly in this room, and I didn't know Harmon.ie was sponsoring that session just an hour ago. We have as part of our digital workplace program, where we embark on adopting the latest technologies, a strong journey on the Microsoft road with respect to Exchange, SharePoint, and other platforms. And enabling simpler access and adoption to these technologies is highly supported by Harmon.ie. So, how you can share documents in a simple way, how you can interact with people. Even on my mobile phone, I can get access to all my documents wherever I am in the world thanks to our Harmon.ie app.

Michael Krigsman: So, the …

Rainer Baumann: So, it even sounded like a commercial now.

Michael Krigsman: [Laughter] Well! So, the key is the simplicity of managing the information, which sounds like it's an essential part of your broader strategy for the business.

Rainer Baumann: It is. In fact, if you think about it, our business is so diverse. So, we have experts that focus on marine cargo hull ─ meaning on the hull of a ship [carrying] cargo. You have a few folks focusing on this with respect to documents, pricing, and a lot of knowledge. And, we have hundreds of thousands of such specialized areas. They need to manage their knowledge somehow and evolve it. Even though we rely highly on the individual experts, naturally, a company like us also wants to encode some of this knowledge to sustain this [expertise] beyond just the individuals upon which we rely heavily.

Michael Krigsman: Fantastic! Well, this has been an absorbing discussion about an industry that is in the midst of dramatic change. We have been talking with Rainer Baumann, who is the CIO for Information at the large, Swiss insurer called Swiss Re. Rainer, thank you for spending time with us today! It’s been a lot of fun! And everybody, I hope you will come back on Friday, because we have another installment of CxOTalk, and we’ll see you then. And, thanks again to Harmon.ie for sponsoring this episode. Bye-bye!