Building Digital Insurance Products - how hard can that be?

Spoiler alert: Quite hard.

Where do we start? Munich Re has been around for almost 140 years – the average life span of a Fortune 500, multinational corporation is only about 40-50 years. So the company has adapted to the changing of the world over the last decades, which implies there is at least a little bit of innovation going on. The power of the brand and the geographic reach are fantastic, the core business is doing really well.  

Recently, Munich Re is generally perceived to be on the forefront of innovation and digitalisation in insurance (you want data on that? Good on you! See for example here and here).

I have been working on building digital insurance products, in this corporate environment, for almost three years now - first in the Asia Pacific region and now in Germany. Cyber risk assessment, life insurance underwriting, claims management and more: while pushing through many projects over the last few years, I have encountered and learned a few things (views are my own and based on my experience – your mileage may vary).

Risk is a business

Innovation requires risk taking – that’s something that resonates across all the books, lectures, conferences, TED talks and Holi colour runs on innovation. However, the insurance industry is usually perceived to be this slightly stuffy, unsexy and risk averse industry that has a hard time innovating. In my experience, that’s not generally true.

Insurers are per se not risk averse – in contrast, it is their core business to take on risk.

The big BUT here is that the industry in general only wants to take on well-understood risks.

In the innovation domain, that attitude can be a problem.

Everyone subscribes to the ‘fail fast’ paradigm on paper (we have all read The Lean Startup book). However, before a new project can start off, a lot of people start to chime in wanting to understand the risks and associated mitigation measures. In many cases, the discussions on potential risks and outcomes can take longer than the actual project execution! Such a slow start with committees, internal units and executive updates can be demotivating to the project team.

Can you do something about it? I found it helps to acknowledge that this is deeply ingrained corporate culture – you can’t ignore it, or fight it. Instead, it is better to give everyone space to voice their opinions and concerns (but time box and moderate the exercise!). Over time, with more and more (hopefully) successful projects on the way, it will become easier.

New clients and needs

Innovation and digital products are usually not simple extensions of the core insurance business model. Suddenly, there are new users, new markets and new demands that need to be understood. That sounds almost too trivial to write down. However, for a successful B2B company, the concept of talking to clients and end consumers about their new needs is not a daily routine (client managers will disagree with me here).

 In the early days, we have developed some of our new digital offerings without enough market and client validation.

We assumed that since we had been in the traditional business for so long, we would know what our clients wanted in the new digital world. We paid dearly for that attitude.

While the solutions were technologically advanced, the offerings didn’t get enough traction because they did not solve important enough problems.

Thankfully, we have learned from that. These days, we explore a lot more before the first developer writes a line of code, and the products are much more focused.

 It took us a while to identify and build up the right digital design skills in the organisation, though.

Technology matters

I have sat through a few discussions where the idea of Munich Re as a future technology company was played around with. I would not subscribe to that – it will be a technology-applying company, I am convinced, but I doubt we will build the next Big Data distributed file system and sell it.

In many (though not all) use cases in the digitalisation space that I have encountered, you do not need to build a rocket ship that can fly to Mars. You need to be state-of-the-art, yes – but not much more.

A lot of the cases I have seen follow a certain pattern: you digitize a workflow consistently, which allows you to gather data consistently, which allows you to do in-depth reporting and analytics, which in turn provides you with greater insight into the risks and user behavior, and that is a starting point for AI supported decision making. If your use case starts from Excel and email or (worse) paper and phone calls, jumping straight to advanced AI is too grand of an ambition in most cases.

Obviously, you want to make use of modern cloud development capabilities, APIs and SaaS wherever you can (and we do), but executing well on that is such a huge step forward in many cases that it is more than enough. 99% of the insurance digitalization use cases will not benefit from quantum computing or blockchain.

But in a few rare cases you do need to go to Mars, and then you go big.  

In summary:

• Acknowledge people want to understand the risks (after all, they work in insurance)
• Talk to users - don't assume you know them
• Be technologically state-of-the-art, don't overdo it.

(To be continued…)