logo
Beyond HumanBig PictureCatalystsConnected WorldExchangeMarketing MixNew MoneyNew SchoolPeople SciencePulse

Zurich: The Insurance Giant Making Big Claims for AI

Zurich: The Insurance Giant Making Big Claims for AI

insurance insurance

AI speeds up the boring bits of insurance but will it ever displace humans completely? Monika Schulze, head of customer and digital experience at Zurich, answers.

In many ways, insurance is a profoundly human process.

When people buy policies, they do so to mitigate the impact of very personal challenges such as illness, unemployment, accidents and bereavement.

It takes time to explain the nuances of insurance – especially long-term schemes such as life cover. That’s why, when people buy these policies, they mostly do so from human agents.

So how can artificial intelligence (AI) possibly help the industry?

Well, a ‘robot’ might never be able to explain the benefits of a complex policy as clearly as a person.

But when it comes to the administrative process of processing premiums, claims, reporting and pay-outs, the reverse is true.

Zurich Insurance knows this. Its recent test project proved it.

“We ran a pilot in the UK, which looked at medical claims reports,” says Monika Schulze, head of customer and digital at Zurich Insurance.

“We found it took a human 58 minutes to process the claims. With AI, we reduced that to three minutes. We had fewer mistakes, it was quicker and it helped us to increase the quality of claims and reporting significantly.”

In fact, this is just one pilot among many. In another, Zurich used AI to reduce personal injury claims processing time from an hour to just five seconds. The company told Reuters this saved 40,000 work hours.

And in March it completed a six week test in which it processed non-emergency home and motor claims using an AI-powered chat bot called Zara. ‘She’ could log the policyholder’s initial query at any hour, seven days a week, before passing it to a human claims handler.

Clearly such AI-based efficiencies can deliver transformative cost savings for insurers. A 2026 white paper by LexisNexis concluded a one per cent improvement in loss ratio for a £1 billion insurer, with better use of data, can be worth £10 million on the bottom line.

Machine Learning and Artificial Intelligence in Insurance

However, Schulze is careful to point out that it’s not enough to view AI merely as a tool to save time and money. Rather, the technology has to serve a broader goal of customer satisfaction.

“I think there’s a clear link between NPS (Net Promoter Score) and customer lifetime value,” she says.

“The more satisfied the customer is the higher the lifetime value is, because they stay longer and they promote you. You can use techniques like AI to make a customer’s life simpler, and of course this is the link between the two.

“But it’s just one piece of puzzle. You have be very clear about the customer experience first and then decide what you need to do to increase it.”

That’s why Zurich is firmly committed to a model that places AI and digital strategies in a blend with telephone-based call centers and face-to-face human agents.

“There’s a triangle, which I see as extremely valuable for customers,” says Schulze. “You have to give people the choice. You can’t just say everything is digital, and equally you can’t say you have to sit on a sofa with an agent. The trick is to combine them and communicate what other options are available. Then you can adapt over time according to what customers do.”

Schulze adds that this ‘triangle’ approach lay behind a recent launch of a digital platform for small businesses in Spain.

The scheme, powered by US-based CoverWallet, lets SMEs quickly analyze their insurance needs, and find the right product for their business.

Zurich Launches Digital Platform for Small Business

“The project in Spain has exactly that [triangle] model. There’s a front-end online digital tool, which is extremely valuable for small business partners. They can do everything online if they want, or they can get in touch with an agent or call if needed.”

Ultimately, Schulze argues, the aim of all these innovations and tests is to drive towards services that treat customers as individuals.

“When you look at online, you have to identify certain questions,” she says. “Then you can adapt to the needs of one person. You don’t want to address everyone in the same way. You want to be as personalized as possible – and there AI can help a lot.”

CHANNELS