I’m thrilled to sit down with Simon Glairy, a renowned expert in insurance and Insurtech, whose deep knowledge of risk management and AI-driven risk assessment has made him a leading voice in the industry. Today, we’re diving into the groundbreaking launch of Unified Risk Reasoning, a game-changing advancement in digital risk processing for insurance professionals. In our conversation, Simon sheds light on how this innovation transforms traditional workflows, the power of AI in mimicking human decision-making, and the practical benefits for insurers, brokers, and reinsurers. We’ll also explore the platform’s unique features, from multi-source data integration to automated web research, and discuss the future of automation in the insurance space.
How does Unified Risk Reasoning stand out as a breakthrough for the insurance industry, and what core problem does it solve?
Unified Risk Reasoning is a significant leap forward because it tackles a persistent challenge in insurance—fully automating the pre-decisioning workflow. Historically, risk automation has been incomplete, leaving insurers to rely on manual processes for reviewing and finalizing data. This new approach acts as a proactive partner, not just gathering data but reasoning through it to finalize critical fields for risk decisions. It’s a big deal because it streamlines the entire process, reducing human intervention and paving the way for true end-to-end automation.
What sets Unified Risk Reasoning apart from the traditional risk automation tools insurers have relied on in the past?
Traditional tools were mostly about data collection—pulling in information but stopping short of interpreting or deciding on it. They often left gaps, requiring manual review to reconcile conflicting data or fill in missing pieces. Unified Risk Reasoning goes further by adding a controllable reasoning layer. It evaluates data from various sources, resolves discrepancies, and makes decisions on final field values, effectively bridging the gap between raw data and actionable insights.
You’ve described this technology as thinking and acting like a human. Can you unpack what that looks like in a real-world insurance scenario?
Absolutely. When I say it thinks and acts like a human, I mean it mirrors the way a seasoned underwriter would approach a risk assessment. For instance, imagine a submission comes in with incomplete property details. A human might cross-check the submission against public records or web data, weigh the reliability of each source, and decide on the most accurate value. Unified Risk Reasoning does exactly that—it assesses multiple data points, contextualizes them, and arrives at a reasoned conclusion, all while documenting its thought process for transparency.
What kind of changes can insurers, brokers, and reinsurers expect in their daily operations with this technology?
The impact is transformative. For one, pre-decision workflows will speed up dramatically since manual reviews are minimized. Underwriters can focus on high-value tasks rather than data reconciliation. Efficiency gains are huge—think faster processing of submissions, renewals, and claims. On the cost side, reducing manual labor means lower operational expenses. Overall, it frees up teams to prioritize strategic decision-making over routine grunt work.
Can you explain how the multi-source reasoning feature works and why it’s so powerful for risk assessment?
Multi-source reasoning allows the platform to pull data from a mix of places—submissions, external databases, and even web research—and combine them intelligently. What’s powerful is how it interprets these sources as a whole, not in isolation. For example, web data might clarify a vague submission detail, adding context and reliability. By configuring these combinations in natural language, users can customize how data is prioritized and finalized, ensuring straight-through processing with minimal human input.
The platform includes a comprehensive data library. How user-friendly is it to integrate new external data sources into the system?
It’s designed to be incredibly user-friendly. You don’t need to be a tech expert to add new data sources. Users can define and connect external sources using simple, natural language instructions, and the platform handles the rest. This accessibility means teams can continuously expand their data pool, which the system then uses to refine incomplete or incorrect fields by cross-referencing against the broader risk or claim context, ensuring a more accurate picture.
I’m curious about the agentic web research capability. How does it assist with finding risk data online, and can you share a practical example?
This feature automates what used to be a time-consuming task—searching the web for missing risk information. The system conducts targeted searches, pulls relevant data, and applies it to the submission. Take, for example, an underwriter assessing a commercial property with limited details. Instead of spending hours googling ownership records or recent incidents, the system does it instantly, retrieving and integrating that information into the risk profile. It’s a massive time-saver for underwriting and claims teams alike.
What are automated enrichment chains, and how do they contribute to building a more detailed risk profile?
Automated enrichment chains are about creating a domino effect with data. The output of one enriched field becomes the input for another, building a layered, comprehensive risk profile. The system automatically figures out the sequence—understanding which data points depend on others and in what order they should be gathered. This enables advanced workflows where users can progressively refine data through multiple sources, resulting in a much richer and more accurate understanding of the risk.
Transparency is emphasized through the Chain of Thought feature. How does this help insurance professionals trust and act on the system’s outputs?
Transparency is critical for adoption. The Chain of Thought feature breaks down the system’s reasoning in plain, natural language, so underwriters and claims handlers can see exactly how a decision was reached. It summarizes the key details and logic behind each step, making the output not just a black box result but a clear, actionable insight. This builds trust, as professionals can review the process and feel confident in the system’s conclusions.
Looking ahead, what’s your forecast for the role of AI-driven automation in the future of insurance workflows?
I believe AI-driven automation, like what we see with Unified Risk Reasoning, will become the backbone of insurance workflows in the next five to ten years. We’re moving toward a world where manual processes are the exception, not the norm. AI will handle increasingly complex tasks—beyond data processing to strategic risk analysis—while integrating seamlessly with human expertise. The result will be an industry that’s faster, more cost-effective, and better equipped to handle emerging risks in a dynamic global market.
