Caroline Dautzenberg on the use of AI within PGGM Pension Services

Caroline Dautzenberg, Director of Patricipants in Pension Services is closely involved in advancing the use of AI within PGGM. She sees many opportunities to use AI to deliver smarter and more tailored services. ‘Within our customer contact organisation, we immediately saw ways to help participants and employers more quickly and effectively,’ she says. 

Caroline is responsible for Pensioenfonds Zorg and Welzijn's participant journeys, an approach that starts with the needs and experiences of our clients’ participants when designing services, helping us to support them better during key life events on pension matters. From her role in the Pension 

The rise of AI 

When AI tools first appeared, Caroline was especially surprised by how quickly the technology became accessible. She started experimenting with ChatGPT and later with the PGGM Private Copilot. ‘Simply by trying things,’ she says. ‘I also really benefited from the Prompt Engineering course. I have now created my first agents in Copilot Studio and plan to explore that further,’ she adds with a smile.’ 

Opportunities within Pension Services 

It soon became clear AI had great potential within her unit, particularly in the Customer Contact Organisation. ‘We immediately saw the potential of speech-to-text technology to record conversations. This reduces the time spent on administrative work, allowing colleagues to focus more on helping participants and employers. With the help of AI, and by continually refining our prompts, we can now capture high-quality summaries of conversations,’ Caroline explains.  

The next step involves using this data to improve client reporting. ‘In the past we experimented with emotion recognition, but the technology was not quite ready back then,’ she says. ‘Now we are working on a writing assistant that can respond to the customer’s emotions.’ 

The AI strategy of Pension Services 

The AI strategy within Pension Services is built around four pillars: client interaction, efficient processes, employees who integrate AI into their daily work, and IT capability. For each pillar a vision, a moonshot, has been defined, with a clear growth path towards it. 

Beneath these pillars sit more than 300 use cases, clustered and prioritised. ‘There’s no shortage of ideas,’ says Caroline. ‘The real challenge is finding the right balance between speed and control. We need to comply with regulations, work within a standardised and flexible management process, ensure the availability of data and a solid IT architecture, and clearly define roles and responsibilities. And of course, we must invest in the right skills and culture; AI literacy is crucial.’ 

Fortunately, Pension Services can build on its experience with implementing the data strategy. 

AI as part of Future Ready 

Earlier this year, Pension Services launched its AI strategy with the AI Awareness Kick-off. ‘We organised an interactive AI Parade featuring an AI Photobooth, Magic Mirror, Sketch Synthesiser and VR headsets. At the AI Marketplace, colleagues gave live demonstrations of applications already in use, from smart summaries to chatbot demos.’ 

The kick-off fits perfectly with the Future Ready way of working at Pension Services, where learning through experience takes centre stage. She explains: ‘We stimulate the senses, learn by doing, seeing and hearing. This way, we continue to grow and stay ready for the future.’  

In September, the initiative continued with the Future Ready Festival, which also focused on AI literacy. With nearly 30 workshops, more than 1,500 participants and an average rating of 8.5, the event was a great success. 

From experiment to scale-up 

Several successful applications have now emerged from the collaboration with Innovation. ‘For all funds, there now is an AI assistant that can quickly answer questions about the new pension scheme. We also use writing assistants that automatically convert content to B1 level, following each fund’s writing guide. In addition, a compliance tool has been developed to help identify employers who may fall under the mandatory participation scheme,’ Caroline says.  

A new addition is Rosa, an AI chatbot that supports the Internal Control team by analysing the root causes of incidents and recording them in Cerrix (the system for internal control and risk management). ‘How cool is that!’ Caroline concludes enthusiastically.