Bernd Oerthle helps a German top 10 insurance company drive satisfaction with AI
Based in Stuttgart, Germany, Bernd Oerthle leads the business analytics team for one of Germany’s largest insurance companies, Alte Leipziger – Hallesche. His is a large team — 30 people — but with over a decade using IBM® Cognos Analytics Bernd thrives in his role.
Oerthle consolidated two separate analytics units into a single team in July 2018. “The two teams were doing things in different ways, with different tools, and only serving their business group — there was no company-wide vision,” says Oerthle. “We had three different versions of Cognos running.” The consolidated analytics team performed strategic research to rethink how to provide information to internal customers. Oerthle’s ultimate goal is to enable his customers to work with their own data. Rather than being the team rushing to deliver reports, he wants to give his customers the tools and knowledge to quickly inform their business decisions by themselves.
Delivering day-to-day value and strategic insight
Bernd Oerthle and his team use Cognos to deliver reports to all operational units of the company — life and health insurance, property insurance, and other financial services products. “Our reports deliver sales numbers and customer information, and also help drive strategy around new products in every division,” says Bernd.
Oerthle and his team have developed about 250 operational use cases to help employees get the data they need to make better decisions. “One of our most popular use cases is called Statistikdatenbank, which helps users calculate new tariffs in health insurance,” he says. Other use cases help employees see damage histories for specific customers and adhere to legal requirements around commissions.
The company also uses Cognos to improve internal performance and employee satisfaction, revealing whether employees are overloaded with work, or where specialists who can address different types of questions when necessary. “We can also set and track target achievements for our sales department,” he says.
Day by day, Oerthle helps guide employees to the answers they need. “We’ve been using Cognos for many years, but only about 40% are regular users,” says Bernd. “The other 60% just know they need answers. They need help finding BI solutions to make simple improvements. And I’m known as the guy who knows the queries, or has the data, or at least, the guy who can find what you need,” he laughs.
Leveraging data to make better predictions
Bernd Oerthle sees himself as a data scientist — he’s always looking for innovative ways to build new models that draw upon the wealth of the company’s data. By integrating Jupyter Notebooks into Cognos, he can use machine learning to leverage the data work from previous projects. “For example, we’re currently developing a prediction model for savings in hospital billing that will reduce manual invoice verification, reduce errors, and improve customer satisfaction,” he says.
Cognos Analytics serves as the data storage system, and the model is developed, calculated and validated in the Jupyter notebook using Python. The results are then displayed directly in a Cognos report. Thanks to the existing interface, the data is transferred without manual effort. Oerthle says, “I see this as an ‘appetizer’ for our customers as we move further into AI, using Watson Studio and Watson Machine Learning.”
Exploring machine learning and the AI Assistant
Regarding what’s next on the analytics roadmap, Bernd Oerthle says his next project will involve predicting insurance policy cancellations. “We’re using machine learning algorithms to develop more accurate predictive values for the probability of a termination.”
“In all of our predictive projects, we gain insights about our customers and partner service providers that would be impossible without data-driven models,” he says. With this new information, he is able to gather new data sets including “classification algorithms or shopping basket analyses, [which] can discover important ‘constellations’ in the data that our experts couldn’t easily see.”
Bernd Oerthle is excited about the future of Cognos at Alte Leipziger – Hallesche. “When our test users started training on the Explorer and AI features, they really loved them, especially the AI Assistant,” he says. “So we’re looking forward to rolling out those features in midf-2020.”
Being an Analytics Hero
“This was very unexpected, and I feel honored to be nominated,” he says. “But it was only possible because of the team, who stays focused on our objectives every day. They put so much time and enthusiasm into improving our employees’ lives and providing the best service to our customers.”
Oerthle knows that success begins with talking to customers. “If you don’t understand what people want, you’ll fail,” he says. “Sometimes they want a fancy visualization, but often they just want access to their data.” Perception of the analytics team is also important. “I don’t want them to see us as an ‘evil empire’ who hoards all the data,” he laughs. “We stress that we’re a development unit who is here to make their work more efficient, faster, and better.”
Every day that Bernd Oerthle’s customers can do their work using his solutions and be self-sufficient is a success. He adds, “We want things to run fast, stable, and smooth, even if they don’t realize that we made it all happen.”