New digital trends in Silicon Valley – the pace has accelerated

Silicon Valley is and remains the driving force for trends in the digital world. The latest trends include networked AI systems which will solve problems of a complexity that we can scarcely imagine today. During my most recent visit to California, however, it was primarily about “today”: What technologies can already help us at ERGO right now?

In 2020, as part of ESP2, we set ourselves ambitious targets. One of them was: By the end of 2025, we want to be the digital leader in the insurance sector in Germany and our core international markets – and remain so in the long term. We now manifest our ambition in more than 110 Group-wide applications with artificial intelligence (AI). Thanks to our more than 80 start-up cooperations, we are continuously digitalising our work processes – not only through AI but also by means of process mining, robotics or phonebots. Interaction with our customers is also becoming increasingly more digital or “more hybrid”.

What new technologies could be useful for ERGO?

But we’re not resting on our laurels, because the rate at which new AI applications in particular are coming onto the market has increased yet again in 2025. We therefore want to know early on what new technologies could be useful for ERGO. This makes the work of our innovation scouts, who keep an eye on developments worldwide, so important.

As a company, we benefit from the growth opportunities of new digital technologies and ecosystems. We can become more efficient and more competitive if we continuously and consistently digitalise our services and processes along the entire value chain. This doesn’t only benefit our customers. As insurers, we make a significant contribution to social, individual and economic stability – especially in globally uncertain times.

We are therefore regularly in Silicon Valley to learn first-hand about the latest innovations.

Google founder Eric Schmidt: “AI is underhyped”

Before my visit, I looked at a talk that Google’s former CEO, Eric Schmidt, had given to health professionals in April – a good introduction to the current topics in Silicon Valley. At the end of the talk, Schmidt asks why there still isn’t a digital model of the human cell. This is meant rhetorically, and he gives the answer straight away: Because it’s extremely difficult. Now though, we’re on the brink of doing this – thanks to generative AI (GenAI).

In his talk, Schmidt therefore says that “AI is underhyped” because we underestimate its potential. In the future, AI will be able to do things that we don’t yet have words for today. “We don’t know where the limits of knowledge lie”, he says. And therefore we also can’t say what will happen when each of us has the “equivalent of the most intelligent person in their trouser pocket”. Schmidt cannot be accused of painting too rosy a picture of the tech world here – he is just stating the facts. But the enormous pace of innovation itself provides the facts that should open the door to an almost unbelievable modern age.

Thus two San Francisco companies – Google and AI expert Anthropic – have built algorithms that will allow AI to be networked. Then we will no longer have to choose between one model or another, because different systems will work together – like an assembly line of AI agents. Together they will solve problems or provide services of a complexity that we can scarcely imagine today. The whole of Silicon Valley is currently talking about such “protocols”. A quantum leap that’s only a few weeks old – based on a technology that’s a few months old. It’s no exaggeration if I write that the pace of development here too is breathtakingly fast (if you would like to learn more about this topic, I recommend this article on the principles of AI agents in our //radar-magazine).

Ready-made AI solutions for almost all areas of ERGO

We could have looked at hundreds of applications during our visit. However, we limited ourselves mainly to those that are ‘enterprise ready.’ In other words, solutions that could be implemented immediately by large companies such as ERGO. I found the following companies particularly exciting:

Anthropic wants to use large language models (LLMs) to significantly reduce AI hallucinations. Its own AI assistant, “Claude”, apparently already delivers more accurate and more plausible results.

Foundation, on the other hand, is using AI to automate policy management and claims settlement at insurance companies.

Hawso offers synthetic “twin” datasets with the same statistical characteristics as the original data, but without real-world confidential or personal information. This allows calculations and simulations to be carried out without any individuals being identifiable.

Mechanical Orchard has developed an AI platform that allows old systems to be migrated to modern cloud formats. What’s particularly impressive is that the AI fixes programming errors by itself. Incidentally, we are also shareholders in Mechanical Orchard via the ERGO Corporate Venture Fund – our investment fund for technology companies.

We also got to know AI tools for HR and marketing which perform valuable support functions for the recruitment of new staff and with online marketing, for example. Whereas in 2024 we spoke about generative AI in general terms, one year later we are now already significantly further on the technical side of things. For each of our classes of insurance, there are now AI applications that could help us too with automation.

Innovations of “tomorrow” – smart glasses that respond to gestures and movements

Talking of “could”, all the solutions I've described have, as mentioned, have reached market maturity. Together with the user departments, we are currently examining which solutions offer added value for ERGO – and how we can quickly integrate them into the ERGO system landscape.

Other technologies, on the other hand, are still a long way off. “AI agent” was thus a magic word that came up at all of our meetings. Now we already do have agents, for example ones that can make restaurant reservations for us. However, as already mentioned above, in the future an entire armada of AI agents will manage highly complex tasks autonomously.

Here is an example:

Tested at Meta: the smart glasses ‘Orion’

The same was true of the Orion smart glasses we tested. It may be some time before Meta’s latest model is available in shops. The metaverse hasn’t caught on in the consumer sector the way that some experts predicted (the situation is different in the case of industrial applications for medicine and technical remote maintenance). Meta’s Orion smart glasses are more similar to conventional glasses than the virtual reality models we have today.

The aim that Meta associates with these glasses is “contextual support”. For example, you look in the fridge and the glasses suggest recipes based on what’s in there. More useful, in my view, are video calls on the glasses, where other videos or graphics which would be too small for a hand-held display could be displayed in parallel. A special feature of the glasses is that they can spot natural gestures and muscle movements and respond to them.

Why, as insurers, should we be concerned with the three-dimensional internet and smart glasses? Because we know that future generations will become increasingly more familiar with such virtual worlds.

My conclusion from my visit to Silicon Valley – and what it means for us at ERGO

First – the pace at which innovations are being developed has increased rapidly. A new “turn of the crank”, as developers call major innovations, happens almost weekly. We at ERGO should therefore remain inquisitive and want to try out new things – as many of us are doing extremely well with ERGO GPT.

Second – despite future AI agent assembly lines, the order will always remain the same: AI supports people. That is the guiding principle, not the other way around. Most contacts in the Valley (including Eric Schmidt) emphasise this! Technology is not an end in itself, it is there to help people. At the end of the day, customers always expect someone who sees them as a person, understands their circumstances, and advises them individually and empathetically. That’s exactly how things are with an insurer – and will also remain so!

Third – AI is not a digitalisation experts’ discipline but lives from interdisciplinary cooperation and will be used to full advantage, especially in the business departments. AI is a discipline of the future which the entire company must get to know and learn about.

I’m now looking forward to the analysis in which we assess the benefits of these technologies for ERGO. I’m convinced that we’ll be using several of the applications that we got to know in Silicon Valley at ERGO in the near future, for in the end what counts now is the implementation.

Kind regards,
Mark Klein

And here are two visual impressions from the trip: