What is needed?

AI in the insurance value chain:

AI can make an insurer faster. But if information exchange with an adviser, service provider or authorised underwriting agent lags behind, the insurance value chain as a whole barely improves or speeds up. The VIP Congress in Leusden reinforced this view: AI’s value is closely linked to data quality and collaboration between organisations. For me, a conference like this is an opportunity to look ahead. What choices will the sector face? And what is needed to turn technological possibilities into better services across the insurance value chain? 

Insurers, authorised underwriting agents, advisers and service providers exchange information every day, supported by standards and shared services. Yet a faster process in one organisation offers little benefit if the handover to the next remains difficult. Meanwhile, the industry faces pressure from a saturated market, further consolidation and strained business models. Legacy IT, vendor dependency, cost control, cyber risks and regulatory pressure make innovation challenging.

The task is to collaborate more intelligently while retaining control over data and responsibilities. This requires better data, modern systems and workable agreements.

Is the information up to date? Where does the data come from? And who is responsible for the outcome? These questions determine whether employees and organisations can trust AI. Data must be available at the right time, and parties must interpret it consistently. Otherwise, an exchange may be technically correct while organisations draw different conclusions. Clear agreements on ownership, access and use are therefore essential. The infrastructure must support these agreements, with attention to security and accountability.

AI only becomes suitable for business-critical use when data, integration, governance and operations function as a whole. The application must have reliable information, fit the workflow and be manageable in daily operation. Responsibilities for reviewing outputs and deciding when human intervention is needed must also be clear.

At the VIP Congress, a Dutch insurer demonstrated how this works in practice. The insurer is building a new data platform while introducing AI applications: a model that predicts water damage to roofs, generative AI that automates the collection of payroll records, and an AI assistant for employees. The platform and applications are deliberately being developed in parallel, with governance embedded from the outset. The message was clear: data fuels AI, and without a sound foundation, applications run into difficulties when you try to scale them.

Organisations that want to use AI in everyday service delivery must establish the necessary conditions together with their value chain partners:

• Ensure critical data is up to date, consistently defined and available.
• Design integration across the value chain, considering the shared process and handover points.
• Define ownership of both data and AI outputs.
• Build in security, auditability and continuity from the outset.
• Organise AI around core business processes, covering development, testing and day-to-day management.

AMIS Conclusion has worked in data-intensive environments for more than 35 years and has extensive experience with data and integration services across the insurance value chain. In practice, we see that alongside technology, standardisation, availability and a shared understanding of data determine whether innovation can truly scale. AMIS Conclusion designs, builds and manages data and integration platforms that make information available to operational processes. We connect data sources, applications and organisations, with attention to security, manageability and continuity. This helps organisations establish a reliable foundation for AI.