Key to achieving this is a high degree of member-oriented digitalisation: Digital solutions allow faster case processing, better resource utilisation and more time for each member, which supports the provision of better member services. They also contribute to keeping costs down, and increased digitalisation thus helps PensionDanmark retain its position as the most cost-effective pension provider in the industry.
PensionDanmark’s member-oriented approach to digitalisation implies that:
- members and companies can carry out all processes and product adjustments digitally on pension.dk in flows that are closely coordinated with Member Services.
- robots function as digital PensionDanmark employees, performing a wide range of advisory and processing tasks that were previously handled by member and corporate advisers, nurses, social workers and other specialists.
- e-mail responses and the dialogue between the chatbot and members are handled by means of machine learning.
- all outgoing messages are personalised based on machine learning/AI and are sent out to relevant members through their preferred channels.
- Members and companies are guaranteed automatic and up-to-date training recommendations on our supplementary training site by means of machine learning based on master data, course history and industry association.
A number of PensionDanmark’s digital solutions, including our supplementary training site, are based on a data exchange collaboration with public authorities.
Data ethics
To be able to provide customer and insurance administration services, we need to access and process relevant data concerning members, companies and employees. Respecting these data and treating them responsibly is fundamental to our way of doing business.
Our approach to data ethics is based on three main principles:
- Transparent and clear information about the use of data and purposes
- Automated and advanced use of data where this benefits our members
- Data security as the cornerstone
Data security as the cornerstone played a key role in establishing and commissioning our new IT platform. The principles on inherent protection of data concerning our members and companies were an integral part of the project right from the determination of solution design and architecture. This means:
- that data are only present in subsystems if strictly necessary
- pseudonymisation throughout the architecture, which means that externally identifiable data are only used in the master data application and have been replaced by internal stakeholder IDs across the rest of the platform
- that user interfaces leave out any unnecessary data
- protection against unauthorised access is incorporated into the infrastructure and data are protected by means of encryption
In addition to embedding data security into our new IT platform, we have developed a number of standards and practices for machine learning and robot technologies to ensure ethical data processing from development through to operations. These standards and practices cover, among other things, access to and selection and storage of model data to prevent discrimination and bias and ensure that unnecessary data are not used.