PGGM improves internal search capabilities across the organization using Elastic Search AI

PGGM is a Dutch pension administrator that manages pensions for 5.8 million participants.* With around 2,000 employees and responsibility for over €250 billion in assets, the organization plays an important role in the financial security of many people in the Netherlands.

PGGM’s innovation department has adopted new technologies to support its diverse workforce and improve efficiency. Using Elastic as the foundation for its search strategy, PGGM moved from older, less accessible systems to a modern, AI-enabled SmartSearch layer that helps employees find information more easily.

Before Elastic was introduced, PGGM employees found it challenging to locate information in the organization’s extensive documentation. According to Lauryn Cuffy, innovation manager at PGGM, the broad structure of the organization and its use of SharePoint made it difficult for employees to find documents.

For example, the legal team often had to search through more than 350,000 documents to find historical cases, policies, or specific emails. To address this, Louic Vermeer, senior data scientist at PGGM, and his team developed a central search engine using Elastic.

The platform brings together different types of data, such as emails, Word documents, and Excel files from SharePoint, into Elasticsearch. The team also used automatic classification and labeling to make it easier for users to filter and find relevant information.

PGGM’s SmartSearch platform uses advanced search capabilities, powered by Elastic. The move from traditional keyword search to hybrid search was a turning point.

This allowed the team to implement a retrieval augmented generation (RAG) workflow, ensuring the AI is grounded in PGGM’s specific data.

The impact on efficiency has been significant:

• Time savings: Different teams reported saving between two and eight hours every week.

• Employee experience: By removing the work of manual searching, employees can focus on more strategic and interesting tasks.

• Onboarding: New employees, who may not yet know the specific company terminology, find it much easier to discover relevant documents using semantic search.

• Data quality: With PGGM’s large volume of data now easily searchable — including archives and older documents — the team can quickly identify incorrect or irrelevant documentation and correct or remove it, ensuring employees find only quality data in their searches.

While the SmartSearch platform is an internal tool, its benefits extend to PGGM's 5.8 million participants. In the customer service center, representatives must quickly find answers within a database of 2,000 Q&As to help participants over the phone.

With Elastic, representatives can provide responses "as fast as possible" while ensuring the information remains factual and consistent. This efficiency ensures that participants receive high-quality support during every interaction.

Looking ahead, the PGGM innovation team plans to evolve the platform from a search engine into a "question-answering engine." By further integrating generative AI and expanding the platform's reach, they aim to provide direct answers rather than just lists of documents.

“The relationship with Elastic is great and it feels like a community,” says Cuffy. “When we have a question, they connect us to other clients who have similar problems to help us solve problems faster.”

*As of January 7, 2024