Skip to main content

Accelerate Time-To-Market with AI

Twitter

Mit dem Laden des Tweets akzeptieren Sie die Datenschutzerklärung von Twitter.
Mehr erfahren

Inhalt laden

With the implementation of our semantic platform semantha, Gothaer Versicherung AG saves a massive amount of time checking broker wordings, new insurer conditions and framework agreements, thus improving the satisfaction of both customers and employees.

With more than 4 million members and a yearly premium income of over 4 billion Euros, Gothaer plays in the Champions League of the large German insurers. Gothaer can convince not only with numbers, they also place great value on high quality personal advice for their customers and partners. Protection concepts are provided in the private sector as well as small-medium enterprises, the self employed and freelancers. With their 200-year history, the insurer again shows their strength – to optimally combine tradition and innovation.

Starting position: Manual work and unsuitable software

Gothaer’s industrial sector, with around 40 employees, serves numerous insurance brokers in the industrial customer sector. Here, brokerages apply their own insurance terms and conditions, on the basis that Gothaer insures the possible risks. These “Wordings” are checked by Gothaer and negotiated individually with the broker.

The terms and conditions of the various brokers often differ only slightly from each other – but have a very large scope. In most cases, they differ only through the formulation of the wordings and structure of the documents. And it is exactly at this point in the process that a high manual effort is required. Until now it didn’t seem that this process could be automated, since text passages with different wordings but a similar meaning would need to be brought together by very simple human knowledge – the theory of meaning – in a time-consuming manner and with lots of care. In particular, the comparison of new text passages with the same or similar passages from the past – although easy to do by hand – is very time consuming, as we can show with the following two examples:

Example Sentence A:
In particular, the interests of the owner or other principals as well as all contractors involved in the contract with the principal, are insured.

Example Sentence B:
The insurance also covers all contractors involved in the contract with the principal, including subcontractors, with their supplies and services respectively. Those considered to be contractors within the meaning of these provisions extend to architects, structural engineers, surveyors, engineers and special experts.

The meaning is the same, but the wording is different. 

Challenge

There are plenty of software solutions which offer a simple text comparison – in most cases they perform a word-for-word comparison or use synonymous terms. Therefore, they can’t offer an advantage here since the text passages are worded very differently.

There is not enough training data for a classical machine learning approach, or a solution would only be useful short term since new data will always be added and the software would need to be continuously retrained.

Solution

The first tests using semantha with example documents from Gothaer provided promising results out-of-the-box which led to an evaluation phase. Since no training data was required and semantha as a platform already has a wide world knowledge, the AI could be directly involved in the work. In close cooperation, the concrete use case and results using real data were analysed.

“semantha accelerates the review of comprehensive insurance terms and conditions. Identical, similar and also differing coverage contents compared to already accepted sets of conditions can be quickly identified”

Christoph Spix, Digital Consultant Composite Industry at Gothaer Allgemeine Versicherung AG

Numerous internal guidelines and previously accepted (or rejected) conditions, as well as text passages from brokers and insurers serve as reference documents in semantha’s database (library). New broker wordings to be checked are now checked with the support of semantha and subsequently negotiated based on the results. The broker then creates a joint version of the conditions, which is again compared using semantha to determine if all passages to be revised have been adapted.

Conclusion

With the use of semantha, the whole process from the broker’s inquiry to the release of the conditions was accelerated considerably. Users at Gothaer talk of an estimated time saving of approximately 40-50% compared to the previous manual processing.

This in turn has an impact on further points, such as a significantly higher processing quality, since the documents are checked consistently. In addition, the customer experience is improved, as the time-to-market is quickened with the faster processing. And at the core, the employee satisfaction is increased, as they can focus their efforts on more complex issues with the newly freed-up time. 

The use of semantha has shown how monotonous work can be eliminated and you can concentrate more on the value-adding tasks. This brings motivation to find further use cases with Gothaer.

The collaboration: exciting and promising for both sides

The collaboration between Gothaer and thingsTHINKING was rewarding for both sides. “We know some companies which deal with NLP (Natural language processing)” summarised Christoph Spix from Gothaer. “But until now we haven’t met any other company whose development is as deeply involved in solving our problems.” Stefan Sebald, Digital Officer at Gothaer, is eager to develop further projects with semantha.

Let semantha convince you!

Customers

We have written extraordinary success stories with some of our clients. You can read them here: Success stories.