How Leading Reinsurer Hannover Re Turns Documents into Organizational Wisdom

Cold Open (00:00-00:15) 

[00:00:05] André Cartschau: 
Documents are the most important thing we have because we do not sell any physical product. We sell safety and security. 

Part 1 – Intro & Welcome (00:00-06:01)

[00:00:16] William McInnes:
Welcome, everybody, to the Enterprise Content Show. I know you’ve missed us terribly. It must have been hard. Franzi, you’re back from a lovely holiday, aren’t you? Where have you been?

[00:00:26] Franziska Thomas:
I was in Mauritius.

[00:00:28] William McInnes:
Mauritius? That is a fancy place to go. Did you have a nice time?

[00:00:32] Franziska Thomas:
Absolutely. Fantastic.

[00:00:34] William McInnes:
That’s good. Did you dream about document intelligence?

[00:00:38] Franziska Thomas:
All the time. There I was on the beach. I was thinking about the lifecycle of document automation.

[00:00:46] William McInnes:
Yeah. So, welcome back to the Enterprise Content Show. I’m Will. I’m still learning that enterprise content can be just as complex as global risk markets, and possibly as unpredictable. And Franzi, how are you?

[00:01:02] Franziska Thomas:
I’m still Franzi, and I’m still here to translate the magic of process, content and AI into something that makes sense for the real world. Hopefully.

[00:01:14] William McInnes:
Really cool. And we’re stepping into the real world today. We’re stepping into the world of reinsurance. This is insurance for insurance companies, like the ultimate safety net. And we’re not just talking about any player in the reinsurance world. We’re talking to one of the biggest players in this market. We’re joined by André Cartschau today from Hannover Re. Please join us on stage, André.

[00:01:51] William McInnes:
Welcome on stage, André. Welcome to the Enterprise Content Show. Really good to see you, André. Thank you so much for joining us.

[00:01:59] André Cartschau:
Thanks for having me.

[00:02:02] William McInnes:
So before we get into organizational wisdom, which is a really juicy theme for us to talk to you about, André, we need to get you warmed up. For anyone who’s never worked in insurance, what is surprising about reinsurance, or what do people tend to be super interested in about reinsurance? Because it’s kind of hidden from popular life, but it’s actually quite interesting, isn’t it?

[00:02:33] André Cartschau:
I brought a real-life example with me, and this will involve you, Will, and Franzi. Reinsurance is a very interesting field of business because, as you mentioned, we are the insurer for the primary insurer. So normally we don’t deal with the car damage that you, Will, may have caused, because that’s the matter of the primary insurer, but we may come into play if Franzi’s 500K Ferrari is affected.

[00:03:01] William McInnes:
Oh, so that’s the thing where you might come into play.

[00:03:06] André Cartschau:
Because the primary insurer may not want to bear this loss, or at least not alone. And let me tell you, there’s nothing that is not covered by primary insurers, and that therefore ends up in the reinsurance book.

[00:03:24] William McInnes:
That is so cool. So you can do something really ambitious and provocative, and then you can cover your downside risks by working with insurance. Interesting.

[00:03:35] Franziska Thomas:
Okay, but by paying for it, right?

[00:03:38] André Cartschau:
For sure. Yeah. You need to pay for your insurance cover, and the insurance company will then look into the risk and say, okay, are we able to bear or handle all those losses on our own, or do we need reinsurance cover as well? And the same goes for reinsurance companies as well. So if you come with an X-million risk to us, we will buy reinsurance for ourselves as well from other reinsurance companies.

[00:04:08] William McInnes:
But what’s the last piece in that chain of events, then? Who’s the one who, in the end, finally carries the risk?

[00:04:16] André Cartschau:
Yeah, I think at the end of the chain are the reinsurance companies. But all the reinsurance companies act more or less as a team. So for sure, we all try to get the good and best pieces of that cake. But at the end of the day, we all need to treat the risks and pay all the claims, and this can only be done together. All huge risks, all huge losses, are always covered by all reinsurance companies.

[00:04:52] Franziska Thomas:
Interesting. And one question on that. How would you describe, or in a word, how decisions are made? Because you said someone is assessing the risk, and then you have to decide what that risk will be worth, and who’s going to make the decision and how.

[00:05:09] André Cartschau:
Mathematics. It’s always about how we estimate the worst loss that could ever happen in this specific area. And then we need to compare that to the offer the primary insurer is offering to us. And it’s a business between humans, so it’s nothing where you can go into a shop and buy reinsurance cover. We need to talk to each other, and sometimes I make the better deal and sometimes you are making the better deal. It’s our strategy to have long-term relationships with our customers, to go in good times and in bad times.

[00:06:01] Franziska Thomas:
I think that’s super interesting because there’s the mathematics layer, and then there’s the human layer. It sounds like it’s an almost perfect fusion of those two things. Really, really interesting.

Part 2 –Organizational Wisdom (06:01-14:18)

[00:06:10] Franziska Thomas:
I think there’s also the content and the context and the history and the previous wisdom. That brings me to our first huge part, learning how you and your organization build that wisdom for the company, and making sure that the people who need the information have the information they need to make the decision.

[00:06:33] William McInnes:
Yeah, that’s absolutely fantastic, Franzi, and I agree. That’s what we want to get into with André. So André, when someone joins a deal or evaluates a risk, what role does historical documentation play in making confident decisions? How do you guys think about that?

[00:06:51] André Cartschau:
We as Hannover Re now have 20 years of document management history behind us. We started with SharePoint 2003 with a connected archive system in the background, but we quickly realised that this did not meet our needs. So in 2009, we started looking for a new DMS and introduced Doxis in 2011. So we are celebrating our 15th anniversary this year.

[00:07:20] André Cartschau:
The biggest challenge was not to place Doxis as a dumb archive at the end of the information food chain. My goal, my wish, is to put it as early as possible in all the respective processes we are dealing with. Because let’s be honest, at some point you have to store and edit the documents somewhere anyway. Of course you can query them on a network drive, but then no one will ever find them again. And you can’t version the documents well with version 0.8, version 0.9, final version and current version in the file name. It’s simply difficult to find the right document. That’s why my approach is always to use Doxis as early as possible because it offers so much added value.

[00:08:12] André Cartschau:
I’ve already talked about version history, but what about annotations? Doxis or SER still jokes about us being the stamp kings. Doxis makes so much possible in this area, and we love our stamps and make really good use of them, especially in accounting or even in sanctions control. And if you are a big fan of folders, because that’s what we’re used to working with on network drives, then you can set them up dynamically in Doxis. We can look at the same file, Will, with a folder structure grouped by the year of signing, but I need the same file grouped by something completely different. So we do have all the reinsurance-related information in Doxis, usually very early in the process, but sometimes also very late because, let’s be honest, not everyone can be convinced. Of course, we also have colleagues who really only use Doxis as a data graveyard at the end of their processes so that auditors don’t complain about it.

[00:09:17] William McInnes:
Super interesting. So on the one hand there’s this idea of a data graveyard, data dump, almost like a dumb repository. But what you’re telling us, André, is that over the 15 years that you’ve been doing this, have you seen how having this historical body of knowledge is enabling people on the front line to be more agile or to have more continuity between projects? What is this data asset enabling for your frontline stakeholders?

[00:10:02] André Cartschau:
In Germany, or in German, we always say the documents are the most important thing we have because we do not sell any physical product. We sell safety and security. So the documents are the most important thing we have as an asset.

[00:10:17] André Cartschau:
You ask how we deal with the entire process, from negotiating until maybe a deal is closed, and we made it, or we try to make it, as easy as possible for our users. You can store all the documents in Doxis as soon as the initial contact is made. If the deal goes through, great, then we generate all the necessary entries in our reinsurance systems, which will result in a file in Doxis, and then the documents can be moved with just two mouse clicks. This shows that Doxis does not stand alone, but is deeply embedded in our processes and integrated in our IT landscape.

[00:11:00] André Cartschau:
An important note: if the deal doesn’t go through, for example because we don’t want to take the risk, then we still have all the information in Doxis. It’s not uncommon for us to see the same risk again in the coming year. Then it’s interesting to see why we didn’t want it last year and who was involved. Perhaps these people are no longer with the company or now have different responsibilities. It doesn’t matter to us because we know where to find these documents. We can exactly say, okay, André and Will were involved last year. What have they talked about, and how was this decision made? This is really, really important for us.

[00:11:42] Franziska Thomas:
So you can ensure that the decisions that were made five, ten, fifteen years ago, and the information that was created at the time, is still accessible to people today if they need to know that information. That brings me to another question. What changes in the process, in collaboration, in working on the contracts or the assessment, when everyone who’s involved actually has the trust and the security that they are accessing the right information?

[00:12:22] André Cartschau:
You mentioned, Franzi, ten or fifteen years, which brings me to another very important topic I would like to add here in this conversation. Because we are not hunters and gatherers, we are of course subject to strict regulation. So we are not allowed to keep these documents forever and ever. We are obliged to delete them once they are no longer needed for our business. We need to take into consideration business purposes and tax issues and so on. If these topics have been clarified, we need to delete those documents.

[00:12:57] André Cartschau:
Try to implement that on your network drive then. Good luck. But in Doxis it’s not a big deal because we have implemented such a deletion workflow, and this workflow automatically checks what needs to be deleted and obtains the deletion approval from the departments involved. If everyone agrees, it is deleted. Are there any objections? Then these must be recorded and evaluated by legal, and next year there will be a new deletion event for this exact process.

[00:13:29] André Cartschau:
So even the boring workflow engine is of big interest to us and very important for us, because we process more than 1,000 workflows per day for accounting, for sanctions control, to involve our claims specialists. So yeah, it’s very important for us. You cannot think about reinsurance business at Hannover Re without Doxis. Everyone knows where to search for the information. For sure, we still have some areas in our company who are in love with the file share, and that’s fine for me. I’m not here to convince. I’m here to offer a better solution in comparison to your file share, and I’m absolutely sure that Doxis is a better solution.

Part 3 – Collaboration, Processes & Adoption (14:18-23:23)

[00:14:18] André Cartschau:
And I offer so many trainings and sessions for you until you have understood why Doxis will make your life easier.

[00:14:26] Franziska Thomas:
I think that is a huge step in the success journey of enabling the users to really see the value and the benefits of Doxis while working with it.

[00:14:35] André Cartschau:
And the good thing, and the funny thing, is that I’m an IT guy. I’m here to operate all the servers and offer trainings and stuff like that, but sometimes it occurs that one business unit affects the other, and one business unit says, okay, watch out, we’re using this kind of workflow because this makes our life so much easier. Why do you not use this kind of workflow? Then for the other business unit, it’s hard to argue against them. This is a very funny situation. We can see it again and again over the past 15 years. It’s not IT who wants to add a new cool fancy feature. Sometimes it’s business to business saying, okay, there is a very good solution available. Why do you not use it?

[00:15:26] William McInnes:
Well, it’s a funny thing. That’s the best thing that can happen. Absolutely, yeah. Let’s move on a bit to one of the next topics that we want to talk to you about, André, and that is the topic of collaboration. We have already touched that a bit, but we want to do a bit deeper dive into how people collaborate, and maybe we can also talk a bit about AI in a bit, but that might be our next chapter. So maybe to give you a leaping-off point, how does document intelligence support collaboration? Because you’ve talked about this being a team sport, both macro and within the firm. Reinsurance companies collaborate, and then of course there’s this collaboration you’ve been talking about. How does document intelligence support collaboration from pre-deal through that lifecycle? What does it enable for people?

[00:16:33] André Cartschau:
Of course, this AI train doesn’t stop at Doxis either here at Hannover, and that’s why we’re now launching a small proof of concept to find out how these new features can help us there. We want to remove the time-consuming and unpopular tasks from the user’s desk. The first thing we are looking at here is the automatic classification of documents. The three of us can look at the same document and still assess it differently, and that’s how it is for us too, especially in an international context. My goal is to standardise this. No matter where in the world a document is saved or originated, I want to have it evaluated and classified according to the same criteria. This is one of our major goals for this small proof of concept. This is not a topic for collaboration. This is to make the underwriters’ and the accountants’ lives easier.

[00:17:27] André Cartschau:
We do collaborate in editing any kind of document within the Doxis system. We are collaborating in sending workflows from an employee to a unit manager, to a division head, or stuff like that. Or if I need some kind of approval for something. But it’s also possible in a negotiation phase to say, okay, Will, watch out, I have a really strange thing on my desk. What do you think about it? So instead of sending emails back and forth, we can use a very easy and straightforward Doxis workflow and collaborate on this specific new risk. Because I, as a trainee, have never heard about that, but I want to gain your expertise on that. So instead of saving the information in your inbox, we try to move it into the Doxis system so that at least the entire team is able to access that information and see, okay, what have André and Will been talking about last year, and why, and how did they come to this conclusion? So this is collaboration in Doxis for Hannover.

[00:18:38] Franziska Thomas:
So that means the most important value for you is the transparency that is added, and a consistency of workflows that are going in the background no matter what. Making sure that the process flows and is not just being stopped somewhere, or that information is being missed or lost in the worst case.

[00:18:59] André Cartschau:
Yeah. If you rely on internal processes and if you rely on the availability and the stability of Doxis, you can trust the system. Trust me. Because for fifteen years, we had no major issue with our Doxis system. I cannot say that in fifteen years there was nothing, but there was no major outage or stuff like that so that anyone here can complain about it. So next to all the cool features you offer within the system, you also provide a very stable and reliable infrastructure.

[00:19:34] Franziska Thomas:
How did you create the processes? Like your team in the background, who designed the master process, kind of? And how much did your internal teams do on their own, and how much support was needed from outside?

[00:19:53] André Cartschau:
We from the IT team, we are a team of four employees being responsible for the entire Doxis thing at Hannover Re, and we can do all the IT stuff. Over the years, for sure, we have gained some business knowledge, but not in a deep dive. So whenever someone wants to have a new feature implemented, we need to talk with the business to ask, hey, what is your pain? How can we help and how can we support you? Then we will discuss and discuss and discuss, create a first draft, create a second draft, and implement and develop, so long as you as the business say, okay, we’re fine with that. This now is a solution which supports our business and fits our needs.

[00:20:45] André Cartschau:
Then for sure, as a regulated company, we have to develop, we have to test, and then we have to transport it into the production environment. Then you are ready to use it. So the normal case is that someone from the business reaches out and says, okay, we do have a problem here. Are you able to help, to support? But sometimes it’s the other way around. Speaking about this AI thing, this was an idea which was developed within our team, where we said, okay, everyone at this company is dealing with this AI stuff. What can we add to this? So we reached out to you, we spoke during the summit, and asked what the next cool thing would be. This was one of the main drivers for initiating this POC now.

[00:21:44] Franziska Thomas:
Yeah, and I assume that in, as you said before too, a very regulated business, it’s probably not super easy to use AI. I assume there are some data protection rules that are more complicated than in maybe other businesses. But on the other hand, it must be practical. So how do you find the best way between what’s fancy nice-to-have and what brings the most value to the business?

[00:22:14] André Cartschau:
It’s always about talking with your own internal customers. We offer all the knowledge we have to the customers, and they may provide all the weird information and all the weird requirements they might have on their desk, and we internally negotiate, okay, what can we achieve with as little effort as possible? It’s like reinsurance business between people. If you have a requirement, reach out and we will try to implement this in the Doxis environment. The only thing which needs to be considered is that it must be usable for all the businesses worldwide. We cannot and will not create a solution which will only fit for Franziska Thomas, because that’s not how a company like Hannover Re can exist. But if your requirements fit one hundred accountants or one hundred underwriters, then thanks a lot for reaching out. This will add additional value to the entire Doxis solution at Hannover Re, and we’re happy to develop and integrate it into our solution.

Part 4 – AI, Use Cases & Closing Thoughts (23:23-29:44)

[00:23:23] Franziska Thomas:
Great. And maybe one last question to finish. André, do you see AI accelerating how organizational wisdom accumulates in Hannover Re, or is it too early to speculate what possibilities exist? When you and the other smart people in the organization think about what happens next, is it clear yet?

[00:23:53] André Cartschau:
I cannot speak for Hannover Re as a whole. I can only speak for my tiny Doxis view, and there I do have some expectations in terms of AI and all the things you might have to offer sooner or later.

[00:24:09] André Cartschau:
One big thing for us is redaction of PII data. We are dealing with almost unlimited PII data here at Hannover Re, and we want and need to get rid of it because it’s very important for us. It’s not uncommon that claims are reported many years later, and if we do not have those documents, we cannot prove what was agreed with the primary insurer. We have seen this happen before by not using Doxis. Despite all the costs, we have even saved a seven-figure sum because we did not have to pay out on claims, thanks to Doxis. So redaction of PII data is very important because then we are not forced to delete those documents, which is, as mentioned earlier, a huge asset for us.

[00:25:08] André Cartschau:
Then we do have some more minor points in terms of AI. It’s kind of superhuman search. This is the vocabulary Doxis uses very often. The thing that I would like to have is a customer relationship summarised for me. I do not want to have a single document summarised, because Adobe can do that too. But if I’m, for example, responsible for a customer named Doxis, then I want to know exactly what contracts do I have with them? What is special? What’s the claim situation? What do I need to pay attention to when visiting this customer? So this is a really big thing for us.

[00:26:08] André Cartschau:
Finally, a very small one, but one that could evolve into a very important topic sooner or later. We want to test whether we can extract content from documents so that we can use this information and process it in other applications. So if we do receive emails or Excel spreadsheets containing claims information, we want to extract information and process it in other internal reinsurance systems, which would make the lives of many employees easier.

[00:26:44] Franziska Thomas:
Yeah, that’s valid. Good points, and a very interesting AI strategy, I would say. You have all aspects of what we, in our mind, think AI can do for companies. That’s the categorisation. I don’t know how many different document types you’re dealing with on a daily basis, but it might be a variety of different document types that need to be categorised. And you said you want to have a standard across different departments and regions, making sure that whether it happens in Germany or in any other country, the category of document type X is always labelled as X in any country.

[00:27:25] Franziska Thomas:
And also extracting certain information, especially from unstructured documents, maybe even handwritten documents or notes, is a very challenging and complex thing to accomplish. So I’m super excited to see how we can deliver that with Doxis, and what your AI experiments and first project attempts will deliver on that. Keep us updated, please.

[00:27:55] André Cartschau:
Will do, will do.

[00:27:57] William McInnes:
Thank you. It’s been fantastic, André. Really super interesting conversation, and we’ve learned a lot. It’s been really interesting understanding that it’s not just about efficiency, although there are definitely efficiency gains. It’s about empowering your people with that context and that confidence, and not just managing those documents, but enabling smart decisions to be made. You’re a decision-making organization, really. So super, super interesting, and we really appreciate you spending some time with us.

[00:28:34] André Cartschau:
Thank you for the interesting conversation, and let’s see what comes next.

[00:28:40] Franziska Thomas:
Absolutely. We look forward to the next chapter.

[00:28:43] William McInnes:
Okay. Great fun. All the best, and take care.

[00:28:48] André Cartschau:
Thank you. Cheers.

[00:28:49] William McInnes:
Well, wasn’t that good? Today we learned that enterprise content isn’t just about files. You already knew that, though. It’s about preserving institutional memory. A super interesting topic, I think, for all of us, given AI and everything. Really interesting. How do we remember what we already know? How is it at our fingertips when we need it? Franzi, what was your take? How was it for you?

[00:29:15] Franziska Thomas:
I do love how he took us through the entire journey, how they started the project, how they still love it, how they enable their entire organization to use Doxis and see the benefit of it. That is very inspiring for me, because I know that is challenging for some companies to achieve that. And I also love how André was talking about Doxis and how proud he is of the project.

[00:29:37] William McInnes:
Yeah. Happy customers make me happy too.

[00:29:40] Franziska Thomas:
Me too, me too.

[00:29:41] William McInnes:
So, really good stuff. Next time we’ll explore how intelligent automation continues to reshape the enterprise, and where AI is taking us next. Franzi, maybe you could finish with the classic YouTuber comments.

[00:29:51] Franziska Thomas:
Yeah, of course. Please subscribe and share and tell us what you’re interested in, what you want to hear next. And remember, keep your content hot, fresh, juicy, connected, and make sure it’s working for you. See you next time.

[00:30:15] William McInnes:
Bye. Thank you, everybody.