Zurich accelerates claims resolution with unique AI solution
Described as possessing the equivalent of 100 years’ worth of claims handler experience, the AI has a reading speed of 10,000 words per microsecond and an average accuracy rating of 98%.
The new technology was trialled between December 2020 and February 2021. It incorporates NLP and knowledge graphs, as well as additional contextual information provided by Zurich to aid decision-making.
In addition to the solution’s improved speed and accuracy over human agents, it allows for greater transparency by including the Ombudsman database and facilitating historic comparison of similar cases.
Delivering better customer outcomes
In order to generate a profound level of subject knowledge, the Zurich-Sprout.ai solution was ‘trained’ using data from 20,000 previous claims. Furthermore, this knowledge isn’t static: the AI’s assessment criteria can be modified with each additional document or piece of data.
According to Zurich itself, this tech development is part of its broader focus on streamlining the customer experience (CX) - a core focus for the post-COVID insurance industry, which is having to adjust to rapidly evolving expectations.
“At Zurich we have been working hard to deliver claims enhancements and innovations that positively impact on our customers’ experience. Our research shows that our customers want speed, clarity and transparency when it comes to making decisions on their claims,” said Amy Brettell, Head of Customer, UK Claims.
“Sprout.ai is a fantastic example of a collaboration which provides a true solution to some of the issues that often stand in the way of creating customer loyalty and helping rebuild trust in our sector. It’s strengthened the customer proposition by cutting settlement decisions to a matter of hours, whilst providing an extra layer of clarity on the outcome.”
Automation and insurance CX
In December 2020, we explored the practical applications and benefits of automation within the claims settlement process. In the roundtable, which included speakers from Munich Re, LexisNexis and Solera, it was concluded that shifting customer expectations would ultimately make investment in automation tech all but inevitable.
Although there is no clear consensus on what the optimal human-to-machine ratio of such solutions should be, the idea that those who do not explore AI in some way would become less competitive was generally agreed upon.
Marcos Malzone, Global VP of Product Marketing at Solera Holdings, commented, “Claims can be settled faster if insurers are able to estimate the extent of losses right at FNOL. For that, they must compel claimants to engage through digital channels and submit damage photos via easy-to-use self-service apps.
“As insurers grow confident in the accuracy of AI, they can automate the approval of estimates up to a certain amount, using appraisers to review the more complex cases.”