Ali Shahkarami, AGCS: Strategic Aspects of Data Analytics (English) – #ID55
What are the responsibilities of a Chef Digital Officer and what are the strategic challenges for the industry from a data analytics perspective? We had a chat with Ali Shahkarami, who is working as CDO at AGCS, about these topics.
In Conversation: Ali Shahkarami, Hartmut Mai and Ansgar Knipschild.
Length: 35 minutes.
Ansgar Knipschild: Hello, and welcome to ID, the podcast for digital industrial insurance. And with me is my Co-Host, Hartmut Mai. Hello Hartmut.
Hartmut Mai: Hi there Ansgar.
Ansgar Knipschild: As you can hear, this is a first for us at ID. It is the first podcast in English. So, I think we will do our very best to improve our school English life in front of you, (laughs) the listeners so be patient. And, as we announced in our last podcast, we want to discuss the topic of data analytics as part of a small series. And that leads us directly to our guest today. It is Ali Shahkarami, Chief Data Officer at AGC and S. Hello Ali.
Ali Shahkarami: Hi Ansgar.
Ansgar Knipschild: Nice to meet you. Ali, you have a background in engineering as I have read, or more precisely in structural engineering. What inspired you to pursue a career in the insurance industry and became the Chief Data Officer of AGC and S? Can you please introduce yourself?
Ali Shahkarami: Absolutely. Thank you very much Ansgar, thanks Hartmut. I am excited to be here. A bit of background on how the journey actually led me to where I am and what you mentioned in terms of structural engineering that is what I did for undergraduate degree that I did. I designed buildings and bridges for a couple of years. But then I went back to school and I continued to drift away from structure engineering, if you will. I did some work in sort of an interdisciplinary area between structural engineering and material science. Then moved on to a small spinoff company from university. They were a handful of us who we were working in aerospace industry, mainly focused on carbon fiber, reinforced polymers. So, all of that, as you can imagine very much engineering focused.
But then I got an opportunity to join a company called Risk Management Solutions. This company creates natural catastrophe models for insurance companies. So, I joined them in developing basically windstorm and earthquake models. And from there Allianz Global Corporate and Specialty offered me an opportunity to move to Munich and lead their research team in national catastrophes at the time CAD risk management team. And I have been here since in a variety of different capacities. Moved on from CAD into innovation, our innovation team, which we called XSE, Cross-functional Smart Evolution. I worked there for a few years and then as of almost two and a half, three years ago, I was tasked to create a central data team which we call ourselves global data office. And I can get into details of that. So, not a straight line from engineering to where I am, but underneath that, it is all using the same tools, same concepts. 2:56
Hartmut Mai: Well, thanks Ali for your introduction. I think that is a fantastic CV you have just been describing to us. You know, I would be interested besides your data journey, right? Which started I think at RMS you mentioned, right? What kind of soft skills did you learn on that journey, which actually helped you to do your job today? What would you say? How would you describe that?
Ali Shahkarami: I think Hartmut, the key for the success for anyone in data and analytics is to understand the business side of the equation and be close to that. So, I mean, traditionally, a lot of the roles that we see on in the technology area and data analytics are very pure technical experts. And more and more, especially now as we are expanding our footprint at AGCS into more use cases and understanding the business and more impact of the business, we realize that this interface between technical experts and business is the key for success.
And then I remember for example, from my background at RMS for instance, there was a lot of client interaction that we had. Not all the modelers were interested in those type of interactions. I was, and a few of other colleagues of mine at the time. And that is interaction on understanding how these tools are used by the clients was key. And we carried that on. If you remember two as well, my CAD days, we had services that we were actually building to enhance the customer experience that we had at AGCS using the tools and expertise that we had on CAD Risk Management. Then we moved on to our XSE days, innovation days. And that was purely built on client interaction and interaction with the business. And now, so short answer to your question Hartmut, the soft skills are basically being able to communicate, speak the same language and understand the challenges that the business experienced. 5:02
Hartmut Mai: So, I mean you have been just describing, you know, a little bit more detail on that journey, which you are taking right now at Allianz. So, from your view, Who is? Is that actually, who is pushing that development? You mentioned the business, you mentioned IT. Who is actually, which function is pushing that? And how do you bring these functions in the end together to all work together on a collaborative basis all geared up to one common goal?
Ali Shahkarami: That is a very good point, because success is not necessarily built by one party. For these solutions, for analytics to really truly transform an organization, there needs to be buy-in and involvement from different functions. So, for instance, our journey now at GDO with everything that we are building with data and analytics, we do not do it ourselves. We have a very close partnership with business, and these are different sort of functions within the insurance value chain, beat underwriting, beat claims, pricing and so on and so forth, these teams, and very, very close partnership with IT.
So, we have created this triangle of partnership between business, us, and IT, and it is very interesting for me that in a lot of interactions I have, some people who are not very close to the group that we are collaborating. If they are on the IT side, they think that we in the data team are a business function, which is a good check mark saying, okay, we are close to the business. And the colleagues on the business side sometimes mistake us with IT. Shows that they see a close collaboration between us and IT. And this is exactly the recipe for success to have not necessarily a competition between different functions or create a solution and try to force people or get buy-in from people to use it, but rather do it jointly together. I think we have managed to achieve that in the last two, three years at AGCS. 7:06
Hartmut Mai: So being a mediary between these two functions. It sounds easy, but I bet it is not always easy. Because as you said, people are looking at you with wearing a different hat all the time. So, talking about that, what are the biggest challenges on your digital transformation journey and data transformation journey in that respect?
Ali Shahkarami: Absolutely. You nailed the problem Hartmut because first of all, I have learned that you cannot turn everybody into a believer. I mean, we still have a lot of discussions around whether we are, or how much impact we have had or how this interaction should work. It has been a journey for us to understand how we should approach this, and what we have managed to do is we have started small with a small group of people who were all driven to create solutions together on the IT side and on the product side with us. And it is interesting that as we succeeded, little by little, this group expanded. And now, we have gotten to a point that we are past the famous tipping points and we have more than we can really respond to in terms of demand coming to this group that we have we are building. And that is the best thing I can ask for.
Has it been easy? No. Have we strike the problem completely and solved it? No. It is still on a daily basis a grind that we have to push the topic. But it was actually interesting, I was talking to one of my team members as we were sort of updating our data strategy for this year and looking back, it has been a transformation over the past two, three years. Given that we also did not start from the beginning, I mean, I talked about our innovation team and a lot of the tools and capabilities and learnings that we have from that team, we carried that forward and we managed to get to where we are today. 9:11
Hartmut Mai: So, I mean, it is one thing to actually look into the future in creating the same data model. So, different departments, different stakeholders can actually start talking the same language if it comes to data. It is another thing actually to look in the past than to integrate data from legacy systems, which does not necessarily actually speak the same language as what you are intending to do going forward. How do you solve these things? 9:36
Ali Shahkarami: Well, this is a problem and again, I hesitate because this is a problem that is all of us in insurance industry are facing. And I do not think when you talk about legacy systems and the landscape, this is anything that is specific to us. It is interesting in the communities that I go, usually talking to other data expert CEOs, everybody is talking about the same problems, data integration, data quality issues, major data quality issues, and having to make decisions based on what we have and so on and so forth. That is not easy. That is a very, very challenging problem and we still sort of are dealing with it.
Having said that, when I started about two and a half years ago, the priorities that I had was to start with the foundation. And the foundation is I want to get my technology stack figured out. I want to make sure that I create momentum around what is everybody in the company should be using and should be driving. That was the first one.
The second piece of the foundation, pillar of the foundation, I would say is data governance. We complete the revamp of our data governance concepts. We have a distributed concept now with data domain owners coming from business, owning those data assets. We give them the guardrails on how they should own and manage these data assets working with us on that and these two enabled us to now start thinking about use cases. Building basically the first versions of dashboards, analytics, use cases and tools with the data that now flow into to our data platform within the governed framework that we have. We are still in that twilight zone in terms of having all our activities focused on this end-to-end concept that we are building, but we have gotten to a point that anything new that comes up has to go within the new concept, at least. 11:46
Ansgar Knipschild: Ali, as we have talked in the beginning before the recording, we both shared that we like podcasts. And this reminds me just what we said of a quote of Richard Harvey. They made a pivot in the last years from a company which offers data for underwriting to now organizing workflows for underwriters and stuff. And the quote I want to cite, I just look at my notes here, “There is not a lack of data holding insurers back too much, but how to handle the data and how to organize data, and this is very hard.” And I think this is what you worked out last answer. And is that what you mean by building the foundation? Beginning by organizing all your data in one common platform?
Ali Shahkarami: No, absolutely. I totally agree with Richard and his comments. You have to keep in mind that insurance is even traditionally built on data. Everything we do is based on actuarial science, and that is very rich in terms of looking at data, historical data, having sophisticated models that actually allows us to assess the risk, price the risk, manage our portfolio and stay profitable while we are protecting our clients. So, data is not anything that is new in our sort of business model.
What is the challenge is that now everybody in the industry, if you are an insurance company, you should be able to do the basics of pricing and portfolio steering and so on and so forth. The problem is we are all dealing with a lot of legacy systems. The problem is that we are not efficient. The problem is that we are not even taking full advantage of the data we have access to. We are sitting on really valuable assets, data, that we can do so much with given the current capabilities that technology offers us. And that is why the attention for all of us at AGCS, Allianz definitely and our peers have moved into how can I create more sort of info knowledge out of the data that we have and improve my processes and provide better decision support so that at the end, our clients have the best experience, the best product that they can, they can have in the market. 14:07
Ansgar Knipschild: And I assume that there is a lot of potential to reduce the frictional costs for handling all the data nowadays. What you send, if you really can organize your data better, if you can give a better accessibility, I think there is really a huge, a very huge potential. If I look at the day-to-day organizations, they now work out in the industry or I think this is a two-digit number and percentage or something like that.
Ali Shahkarami: Definitely. I mean as you mentioned and Richard’s quotes, there is quite a lot of advantages that we can gain by digitizing our processes and letting the data to flow through end-to-end without people having to extract and manually type it in. That is the first problem. And a lot of the issues that we have with data quality is coming from the fact that we are still lagging behind as an industry on this digitalization concept. What I have actually observed in the last two, three years is that there is, there has always been the talk of going in that direction of lose using technology a lot more over the past decade, but I see a lot of acceleration in picking up or investing into these concepts to be able to create more efficiency. So that is one aspect of it.
The second aspect of it is that also the data that we have allows us to be able to predict the future. Everything that we have done in the past is basically looking at our #00:15:31-1# portfolio of risks today, looking at how we want to basically make sure that we are protected on our balance sheets for events that can happen so that we can protect our clients and so on and so forth. But there is a lot we can do to predict the future and help our clients and this is something that now everybody has turned their attention to in our industry. And I think there is a lot of potential for that. 15:56
Hartmut Mai: Yes, I was about to ask exactly that question. So, because every carrier that basically states that they are client focused, that the client is actually in the middle of everything they do, et cetera, et cetera. But if you turn around and actually ask their client, what is your view? And you know, what do you get? What are you getting as a client? What are you getting out of all these data-driven underwriting philosophy, et cetera? Do you see any improvement in services? In fact, do you see any new products coming out of this data stream, which would actually help you to actually manage your side of the business a lot more efficient? And I guess the answer is or I guess that is what we are hearing. There is not an awful lot happening. There is not an awful lot actually getting to the end user, to the customer. So, what is your perspective, or maybe you can actually help us and take us by the hand and look into the crystal ball. What is going to/ what can our clients actually expect you know, on that journey on a short-term basis, let us say in next year, maybe in next three years, maybe next five years? 17:13
Ali Shahkarami: No, definitely the insurance business model is traditionally has been to basically be able to sign a contract and then go away until next year where you sign, renew the contract and go forward. And then if there is of course an adverse event that is impacting our clients, we are there to support them. That is something that in today is world, in my opinion, needs to change. And then some of the things that we did, I mean, Hartmut if you remember, was to engage with the clients to add value to their business with the expertise that we have and with services that potentially we can provide. I would say that as an industry, we are still in our initial stages of being able to actually provide solutions to the clients that creates a lot more interaction, a better product, better experience for our clients. I think we are moving in that direction, but I do not think, in my opinion fast enough.
But going back to your question there, I think there is a lot of potential for solutions in the market. We have a lot of solutions internally that we use that can be turned into something that can provide value to the customers that I remember from the CAD days, for instance, we had a lot of interactions with our customers to give them insights about the risk, something that we use internally to assess their risk and price, their risk and manage them during the natural catastrophes. Why can you not provide that to the client so that they can protect themselves in the case of adverse effect? 18:57
Hartmut Mai: Maybe if you allow me to, to chip in just quickly. Because I would like to get a little bit more concrete in that respect. So, as you know, clients are facing tremendous challenges right now just mentioned three out of many. Aggregation exposure in cyber, for example are leading to the fact that we have a limit crunch in the market right now. So, very much to the detriment of the end use of the client.
Second, business interruption exposures. We have seen a hype of these type of exposures in the/ I would say in the past two to three years. All of the pandemic, same thing, pandemic actually led to the fact that lockdowns were actually closing entire factories in entire countries, having a huge impact on economy and society. I think all of these exposures in the end are all data-driven, and the data is all generated when we think about the data generated throughout the pandemic. What is your vision in terms of using the results of your work, your day-to-day work today in the future to unlock the potential, what insurance can really do for society and for the business with these exposures? 20:24
Ali Shahkarami: The challenge Hartmut, I mean you mentioned risks that are truly challenging risks not only for the societies, but also for the insurance companies. The understanding of the supply chain for companies that like us who cover contingent business interruption and so on and so forth. And the complexity that it has is something that is not a simple nut to crack in that sense. And the answer is not data. That the answer is in solutions that are driven by the data. So, getting to a point that there are mature solutions on these problems that we highlighted takes time. So, for example, cyber you mentioned is one of the risks that there is a lot of, sort of attention around it. There is a good market for it. We still have not gone through a major event for it to mature, like how mature we are for say, hurricanes or earthquakes. And that means that there is still a lot of unknowns in that sense.
But that makes it difficult to be really concrete in terms of what problems can we solve now. I can tell you that we have a lot of data, we have a lot of solutions that we think that are addressing some of these problems, and we try to help our clients on that, on supply chain management on cyber and so on and so forth. But would we be surprised when the next pandemic or the next Thailand flood hits? I am pretty sure there would still be surprises.
In my last eleven years just being with Allianz, we have matured in capturing more and more of events that we think that can impact us and the societies that we live in or operate in. And every time there has been one aspect that has been new, the latest being pandemic. The one after that was supply chain disruptions leading from that pandemic and so on and so forth. So, challenges keep becoming or upping the game in that sense. And I think on the flip side, we are also not as advanced in terms of being able to provide our clients or societies with the solutions that can address them. 22:44
Ansgar Knipschild: One question from my side, Ali, when you talk about data, my understanding was from me, unless you are talking primarily of historic data, perhaps provided by customers or from society or external data sources, whatever. Do you see any advances in getting real-time data from customers? For example, from the industry, from supply chains, from the marine business? Is there really something happening or, because my feeling is just following the news or perhaps talking to one other customer, there is not so much progress and I think there would be a lot of possibilities to really automate things, to make things more efficient. Can you give us some insights here?
Ali Shahkarami: And from what I have seen is that there has been quite a lot of niche products or small-scale activities around, for example, IoT use cases and so on and so forth. There are a couple of new products actually in the last month or so that have entered the market in using data from operations or machinery to be able to predict failure or maintenance, optimizing maintenance costs and include the performance of manufacturers. And you mentioned Marine, there is also telematics in cars. There is quite a lot of activities around those specific products, some of which have matured. But I truly feel that there is quite a lot of opportunity out there to explore.
One challenge is also, I mean, or two challenges, one is that we still do not have as an industry that drive to create new products that are fitted for these to use the real-time data. And the second piece is that from a technology point of view, we are still lagging behind a bit as an insurance industry to be able to actually take advantage of the data that flows in and create intelligence and knowledge around it.
Ansgar Knipschild: But that leads for me to another question because when I follow some conversations of insurers, carriers and customers, I see some very interesting ping pong games there, because on the one hand, for example, the customer says, “Well, your old school economy cannot deliver me digital data for whatever, beginning with invoices or even contracts, very simple things.”. On the other hand, the carrier said, “But you will not give me any data, for example, from your production lines.”
And on the other hand, everybody is talking about industry four zero or whatever. So, it is really interesting to sit there and to watch this. And my question is, is this mainly a technical topic? Because for sure we need infrastructure for that, and how can we exchange data? We need perhaps some neutral data spaces, perhaps something Gaia X is working on or whatever, some neutral place where you can really exchange perhaps anonymized data or whatever. Or is it not even more psychological game or a matter of trust that all the parties, and I do not mention the brokers here, but there are many more parties involved to give data away, or because they have the feeling they lose control. So, we are now going a little bit into data governance, data ownership. What is your impression? How much is the technical thing? How much is it a matter of trust of power perhaps? How do you see that? 26:04
Ali Shahkarami: So, I think that the trust issue is a very relevant challenge. So, and we all have that problem in our daily lives. It is not necessarily insurance companies but you are carrying a cell phone which is a big IoT device measuring a lot of information about you, and sometimes unknowingly you are sharing that with a lot of other vendors that are basically giving you products geared to what, how you behave on your phone.
So, this issue of trust is a major issue. And I think the discussion should hopefully evolve into not necessarily one party giving away data, but collaboration in a way that both parties or all the parties involved in an ecosystem are benefiting from. So if, for example, an organization understands that by creating a product where they provide information about the manufacturing processes, they can become more efficient, more profitable. And at the same time, being covered by product provided by insurance company that allows them to grow, then it becomes more of a partnership and not necessarily, you are the insurance company and I only buy this product for you when I see you next year. So that mode of interaction, the reason that we do not trust each other is because our partnership is really not the same partnership that is needed to be for us to be able to freely exchange data. Now creating central sort of modes of interact/ or exchanging data that can be of a solution. But I truly believe that if there is a product that is created in collaboration with customers that is beneficial to both parties, then the trust organically grows from there.
Hartmut Mai: I think this point is very interesting because in order to design a better trust base with the customer, you need more touchpoints than just an annual renewal with the customer. You know, a bit of negotiation back and forth, premium up, premium down. I am basing my question is now centered around how you started your career as an engineer. Because I think that the engineering capabilities, you know, within an insurance company can actually create exactly more touchpoints if you think about other services built around that risk transfer policy, i.e. risk prevention services, which we sometimes see already in the market but very little how much of these services and you know, maybe an existing one to be developed type of these services can be driven by your work today, i.e., creating data models which actually could flow into these services to help clients mitigate exposures and maybe even prevent exposures.
Ali Shahkarami: No, I think a lot of it. So, I mentioned the fact that everything that we do, we want to turn our focus from being retrospective in everything that we do into predictive and seeing the future. Just imagine, if I can provide that view to our clients about the risk. I mean that is the missing piece right now for them to be able to mitigate the risk, for them to stay in operation a lot more confidently, and also to be able to get a product that at the end gives them even better profitability on their side.
So, there is a value and, and I truly believe that there is quite a lot of room for solutions to be able to actually not only solutions for clients, also insurance companies can also take advantage of a lot of solutions that either vendors or other insurance companies can provide to them. Because the differentiation is not necessarily, again, in being able to assess the risk, but in order to is in providing capability for the clients to be able to see the future and prepare and be more resilient in their operations. What you mentioned on engineering, that is key because that is the specialty that we insurance industry have, the insight we have, and that is directly applicable to our clients. So, being able to roll that out, extend that to our clients, not only helps them, but also reinforces the relationship we have with them. 30:49
Hartmut Mai: So, that would be an extension of the existing business model of a traditional insurance company focused only on risk transfer?
Ali Shahkarami: Definitely, definitely. And I think it is/ we are all coming to a point as an industry that this differentiation would be something that is going to drive the future success of companies because we have been traditional, all of us for a long time. Now, everybody is focusing on digitalization, investment into data analytics and services is just a very logical step after that to be able to reinforce that relationship with the clients and provide more resilience to them services for that.
Hartmut Mai: So, and maybe just the final question for my side, Ansgar? Maybe you actually have something lined up there you know, in your backlog there as well. I would be interested in your view on where your company stands today, benchmarked against the rest of the market and how you would actually judge the market. You know, when, let us describe that journey from zero to a hundred. A hundred means it is all digital in terms of corporate insurance. Where would you rank the market today on that journey? Be it twenty, be it already at eighty and how would you actually benchmark your company against that?
Ali Shahkarami: Marvelous, I am a data guy, so, my estimates need to be based on facts. So, I am going to deviate from that concept and be a bit more qualitative in my answer this time. And I hope that you and also our audience, forgive me for that. But I think from, if I say a hundred is the potential that we have as an industry to be completely embedded into a partnership that is continuous with our clients, we are probably around forty, fifty out of a hundred, whereas there is still a lot of room to reinvent our business model and be client-centric using technology and capabilities around analytics that we have.
The interesting thing is as I mentioned the last few years, I see a lot more drive in that direction. And I think with the speed and the momentum that is being built, I think if we have this conversation in three years, in five years, it would be a completely different picture. Now, the corporate insurer solution providers, AGCS being one of them, of course it is in a spectrum where some of us are good in one aspect, some are good in the other aspects. And I would say, if we look at the pack, I would definitely say that we are in the leading pack. Are we the leading sort of absolutely the leading company on that front? Probably not, but we are a company now that is from our top management perspective, over the last five years, there has been a lot of attention into technology, into transformation, investment into building us to become more and more data-driven and in connection with our clients. So I am excited for what is going to come up in the next few years and I hope that we can continue not only being one of the best in the market, but also pushing the insurance industry in that direction. 34:24
Ansgar Knipschild: So, gentlemen, thank you very much for the inspiring conversation. And Ali, thank you especially for all the insights you have given to us. And yes, I wish you all a good time and say, see you soon. Goodbye.
Hartmut Mai: Thank you, Ali. Thank you as Ansgar.
Ali Shahkarami: Thank you very much.
Ansgar Knipschild: Bye.
Ali Shahkarami: Bye-bye.
Hartmut Mai: Bye.
Der Podcast „Industrieversicherung Digital“ ist eine Initiative für den offenen Austausch über die Digitalisierung von Industrie- und Gewerbeversicherung: Versicherer, Makler, Kunden und IT im direkten Dialog.
Machen Sie mit! Wenn Sie ein spannendes Thema, einen Erfahrungsbericht oder einen persönlichen Standpunkt mit Kolleginnen und Kollegen diskutieren möchten, melden Sie sich bei uns: E-Mail und LinkedIn-Gruppenlink auf der Mitmachen-Seite.
Podcast abonnieren über: