Insurtech's AI Dream Hinges on a Data Reality Check

NEW YORK, NY – June 04, 2026 – The Javits Center was buzzing last week with the promise of an artificially intelligent future for insurance. At Insurtech Insights USA 2026, an event that drew more than 6,000 industry professionals, the consensus was clear: AI’s place in the sector is no longer a question. Yet, beneath the veneer of frontier models and automated agents, a far more mundane and urgent message emerged. The industry’s ambitious AI staircase, as one executive put it, is being built on a foundation of crumbling, fragmented data—and the entire structure is at risk.

“The data foundation has to match the ambition,” Kristoffer Lundberg, CEO of conference organizer Insurtech Insights, stated, capturing the central tension of the event. While keynotes from tech giants like Allianz, Anthropic, and OpenAI painted a vivid picture of an “AI-Defined Insurer,” corridor conversations and technical track sessions repeatedly returned to a less glamorous reality. Before insurers can “dream big or go home,” as Allianz Group CTO Christian Freytag urged, they must first undertake the costly and complex work of cleaning their own house.

The High Cost of a Crumbling Foundation

The gap between AI ambition and data reality is not just a talking point; it's a quantifiable crisis. The core of the problem lies in decades of accumulated legacy systems, where critical information is locked away in unstructured documents, siloed databases, and inconsistent formats. This operational drag is why Gartner estimates that by 2027, a staggering 60% of AI projects in insurance will fail to deliver their expected return on investment, with poor data quality as the primary culprit.

One of the most striking statistics came from a panel on claims transformation, where participants noted that adjusters spend up to 80 percent of their time acting as human “switchboard operators”—manually finding and moving data between disparate systems rather than performing the core functions of their job. This inefficiency is not just a drain on resources; it's a direct hit to the bottom line. A recent whitepaper from the Global Insurance Institute estimated that data errors and their downstream effects cost the global insurance industry over $15 billion annually in rework, compliance fines, and missed opportunities.

In a session titled “Data Foundations to Decision Power,” speakers from reinsurance powerhouses Swiss Re and AXA XL detailed their methodical approach to this challenge. They described mapping AI capabilities across dozens of domains and deliberately prioritizing moderate-risk use cases first. The strategy is to build a track record of successful delivery and internal trust before tackling high-stakes financial and compliance applications. The message was unambiguous: for organizations still wrestling with legacy documents written a decade ago, the architecture is the problem, not the AI.

The 'Superhuman' Adjuster and a New Underwriting Paradigm

While the data challenge is daunting, the potential reward for solving it is transformative. The goal, as articulated across multiple sessions, is not to replace human experts but to augment them. Andy Cohen of Snapsheet framed the ambition as making adjusters “superhuman” by automating the tedium that consumes their day. Free from the role of data courier, claims professionals can return to the work that requires uniquely human skills: empathy, complex judgment, and nuanced investigation.

This shift promises a significant “morale dividend.” Panelists argued that employees engaged in meaningful analytical work, rather than repetitive data entry, are more satisfied, stay with companies longer, and ultimately deliver better results for customers. The same principle applies to underwriting, where AI, powered by clean data, can sift through thousands of data points to flag risks and identify patterns that a human might miss. This allows the underwriter to focus on the complex, outlier cases that define profitability.

Allianz, a leader in this space, has demonstrated the power of this approach. By investing heavily in unifying its data platforms, the global carrier has enabled AI-powered fraud detection systems that have reportedly reduced false positives by 30% and improved detection rates by 15% in some markets, turning a strong data foundation into a tangible competitive advantage.

A New Capital Discipline: Profitability Trumps Hype

The industry's pivot toward foundational work is being mirrored and reinforced by a dramatic shift in the capital markets. The era of “growth at all costs” in the insurtech sector is definitively over. A panel on funding, featuring investors and MGA leaders, reflected a market that has matured from speculative hype to a demand for hard numbers.

“If your hunch is that the right thing is to grow at all costs, we should probably get out of insurance,” stated Matthew Jones of MS Transverse Insurance Group, delivering one of the most direct lines of the conference. Investors, sobered by rising interest rates and a volatile macroeconomic climate, are no longer funding speculative ventures. According to recent reports from CB Insights and PitchBook, median deal sizes have stabilized, and due diligence now rigorously scrutinizes underwriting profitability and clear paths to value accumulation.

There is a pronounced preference for founding teams that blend technological savvy with genuine, hard-won insurance experience. Interestingly, while venture capital has become more disciplined, reinsurance appetite for the insurtech sector has expanded, buoyed by nearly $100 billion in industry earnings over the last three years. This indicates a growing confidence in mature insurtech models that can demonstrate underwriting profitability, with investors favoring flexible structures like ILS and sidecars over the formation of new balance sheets.

Wiring the New Central Nervous System

As the industry grapples with its data problem, a new class of solution providers is emerging to help wire the central nervous system of the modern insurer. The conference’s Insurtech Impact Award was presented to Quantexa, a company recognized for its connected data intelligence platform. The technology works by connecting disparate internal and external data points to create a single, contextualized view of customers and risk, directly addressing the data fragmentation issue.

This ability to operationalize complex, disparate data at scale is the critical missing link for many carriers. Even the most advanced frontier models from companies like Anthropic and OpenAI, whose representatives were prominent speakers, are only as effective as the data they are trained on. Their presence underscored a symbiotic relationship: the AI giants are building the powerful engines, but the insurance industry must supply the high-quality fuel.

The journey is just beginning. Insurers are now faced with the unglamorous but essential task of investing in data governance, modernizing their core architecture, and fostering a culture that treats data not as a byproduct of operations but as its most critical asset.