From data to decisions: Key takeaways from the Head of AI Forum at Insurtech Insights Europe 2026
At Insurtech Insights Europe 2026, the Head of AI Forum brought together senior AI and data leaders from Swiss Re, PZU, Zurich Insurance and Duck Creek Technologies to address one of the industry's most enduring challenges: the gap between AI ambition and production-scale delivery. The session was moderated by Ahmed Al Mubarak, Associate Director, Business Intelligence & Data Science at Howden Re.
“What stood out was the consistency of perspective across organisations of very different sizes and geographies,” said Ahmed. “The gap between AI ambition and execution is not unique to any one firm. It reflects structural characteristics of the insurance industry, and recognising that is the first step toward addressing it.”
The gap between ambition and execution
The panel’s assessment was closely aligned. While meaningful progress has been made, the distance between what is articulated publicly and what is operating at scale in production remains significant. In that context, the most consequential risk any insurer faces right now, the discussion suggested, is overestimating its own maturity.
Where AI is delivering value today
A more pragmatic view of value is beginning to take shape. Claims functions continue to demonstrate earlier and more measurable returns, supported by shorter feedback loops and clearer operational metrics. Underwriting, by contrast, remains a longer-cycle opportunity, where the impact of AI is inherently more difficult to isolate and may take years to fully validate.
The discussion also challenged the narrative that generative AI is displacing established techniques. Rather than being replaced, classical machine learning is becoming more, not less, important within emerging AI architectures. In particular, it continues to provide the structured, explainable outputs required in a highly regulated and actuarially driven industry, often operating alongside language models rather than in competition with them.
Data and regulation shaping the path forward
On data, the panel was unequivocal. The idea that advances in modelling can compensate for weak or fragmented data foundations was described as one of the most damaging misconceptions in the market. Building enterprise-scale data infrastructure takes years of deliberate investment, but it is ultimately what determines whether AI delivers sustained, enterprise-wide value or remains limited to isolated use cases.
Regulation is also beginning to shape these decisions more directly. As expectations around explainability, auditability and control continue to evolve, particularly in Europe, insurers are increasingly required to consider not only what their AI systems can do, but how those systems can be governed and defended.
Building the foundations for next-generation AI
Looking ahead, the conversation turned to the development of more integrated, agent-based AI systems. While the underlying technologies are advancing rapidly, the panel’s view was that organisational readiness, particularly in data, architecture and governance, remains the primary constraint on adoption for most insurers.
“The organisations that will lead on agentic AI are not waiting for a fully defined end-state,” Ahmed noted. “They are building the data, governance and architectural foundations that will allow them to scale as the technology matures.”
The security dimension cannot be separated from this picture. Agentic AI expands the attack surface materially, and sophisticated threat actors are already exploiting this. The panel's position was not that this risk should slow deployment, but that it demands calibrated controls rather than paralysis. Organisations that refuse to deploy AI on security grounds are making a worse trade-off than those that deploy it with appropriate governance.
Distribution and the shifting competitive landscape
The session concluded with a forward-looking perspective on distribution. Conversational AI tools are already beginning to influence how insurance products are discovered and evaluated, acting as a new interface between insurers and customers, with consequences already visible in broker valuations as capital markets price in the disruption to traditional intermediary models. This shift introduces a different form of competitive pressure, not only from within the industry, but from technology platforms that increasingly sit between insurers and end users.
For insurers that have yet to position AI as a strategic priority at the highest levels of the organisation, the discussion pointed to a narrowing window for a measured response.