•
Jun 13
Sand Technologies
In recent years, the insurance industry has significantly expanded its use of AI, with nearly 90% of executives identifying it as a core strategic priority for 2025. But as expectations grow, the focus is shifting. For many insurers, the priority is not just adopting AI, but scaling it across the enterprise to deliver clear and tangible operational and financial results.
This shift was front and center at the recent Insurtech Insights USA 2025 conference, which brought over 6,000 industry players together to explore the future of insurance. The message was consistent across conversations on claims, underwriting and system integration: The industry is shifting from broad transformation goals toward practical, well-defined use cases with measurable returns. Here are the most relevant takeaways for insurers moving from exploration to implementation.
Across the industry, there’s growing recognition that legacy systems carry more than just technical debt. They house proprietary data and logic that are difficult and often unwise to replace. As a result, many carriers are rethinking the idea of full system replacements and instead focusing on adopting a different integration strategy.
Rather than attempting wholesale transformation, insurers are prioritizing targeted AI deployment in high-impact processes like decision-making and risk assessment. By starting with near-term quick wins, insurers can deliver rapid returns while laying the foundation for long-term transformations that allow them to preserve control over critical IP.
Conversations at the event also emphasized a flexible and phased deployment model. While centralized tech oversight is essential to ensure consistency, business teams need the autonomy to move fast. This balance enables AI to unlock outcomes, not introduce new bottlenecks.
Strong governance was discussed as a key enabler in managing legacy systems. Combining portions of native data from legacy systems with new AI-powered applications creates new opportunities. But this requires safeguards to ensure explainability, transparency, risk monitoring and continuous feedback. There’s also an understanding that AI must enhance and not replace human judgment in areas where oversight and accountability are essential.
Claims operations are shifting under pressure from scale, high costs, staff retention and skills shortages and rising expectations. What’s emerging is a redefinition of roles: professionals now focus more on decision-making, while AI handles much of the data aggregation and document processing.
The claims model is also moving from manual data pulling to proactive delivery. In A&H, P&C and L&A, AI systems are starting to surface relevant information, automatically organizing submissions, coding cases and identifying patterns. This allows claims professionals to focus on context, risk and customer communication.
Agentic AI models are gaining traction, particularly in fraud detection and risk transfer. Multimodal agent systems can reason through processes and across structured, semi-structured, and unstructured data, including text, images and audio. They’re already speeding up routine tasks and reducing the administrative burden on frontline staff.
Yet, as several leaders highlighted, the success of these solutions hinges on data access and quality. This is enabled by an emphasis on data preparation, which is empowered by the enhancement of capabilities in aggregation, analysis and augmentation as well. These actions bolster an entity’s efforts to ensure they get the correct data in the right hands at the right time. Building user trust and cultural readiness as it pertains to data readiness is also imperative. Whether it’s demonstrating value, ensuring transparency, mitigating bias or meeting regulatory standards, these foundations are essential to getting the best ROI from AI in claims operations.
Yet to reach
Underwriting is a core insurance function that shapes business performance and customer outcomes. The integration of AI within this function is helping insurers rethink how they assess risk, price more fairly and adapt to changing customer needs. AI and advanced data analytics are enabling more accurate and consistent risk calibration.
Similarly, agentic AI and foundational models are also transforming underwriting. Various insurers shared examples of how these models helped reduce loss ratios, accelerate quote delivery and improve data consistency. Their value was particularly felt in markets facing talent shortages.
However, it was also clear that two factors were shaping the strength of these underwriting solutions. The first is how the data is sourced and organized. Some firms are using the Model Context Protocol (MCP), which enables scalable frameworks for AI deployment. Others are expanding their use of open data sources. Together, these approaches are reshaping how underwriting teams evaluate risk and respond to changing market signals.
The second factor reshaping underwriting in the age of AI is smarter questions. As many underwriting leaders pointed out, having the right data or AI models is insufficient. Insurers need professionals who can bridge the gap between technical expertise and business logic. The success of these AI initiatives lies not just in building better models but in ensuring these models are applied in the right way, to the right problems.
This year’s Insurtech Insights gathering made it clear that the next challenge in insurance isn’t just scaling AI solutions; it’s scaling the mindset. The most successful organizations won’t be those with the most sophisticated models, but those that treat AI as a business capability, not just a technical one.
This shift requires more than investment. It calls for cross-functional alignment, clear ownership and accountability and an increased focus on cultural adoption where human expertise and machine intelligence complement rather than compete with each other. Leaders must prioritize AI’s “last mile”: integration into workflows, user trust and outcomes that map back to real business priorities. Ultimately, the insurers that shift their focus from experimentation to execution, built on strong data foundations, grounded in practical use cases, governed responsibly, will be best positioned to lead through meaningful insights.
Other articles that may interest you
