Today, we’re joined by Simon Glairy, a recognized expert in insurance and Insurtech, to discuss a seismic shift happening within the industry. For decades, critical decisions have been caught in a slow, cumbersome translation process between business leaders, IT specialists, and data scientists. Simon is here to unpack how agentic AI is finally breaking down these silos, creating a new paradigm of agility and strategic control for insurers.
Decision-making in insurance has often been slowed by a translation process between business, IT, and data science teams. Could you share an example of this friction and detail how it tangibly impacted the time-to-market for a new product or pricing strategy?
Absolutely. This friction is something I’ve seen play out time and time again. Imagine a business leader wants to introduce a new discount for homeowners with smart security systems. It sounds simple, but the business team maps it out in spreadsheets, the data science team builds a predictive model in their own environment, and the IT team needs to hard-code it into the core system. It feels like everyone is speaking a different language. The business team’s logic doesn’t directly translate to IT’s code, causing a frustrating, month-long cycle of meetings and revisions. This isn’t a minor hiccup; this kind of organizational drag is responsible for about 70% of the challenges insurers face when trying to scale their AI initiatives, delaying new products and leaving them vulnerable to more agile competitors.
Decision management models were introduced to create a shared language between business and IT. Beyond simply improving communication, what were the most significant operational benefits of this “two-silo” approach, and what specific scaling challenges still remained before agentic AI?
Decision management was a huge leap forward. By creating a unified platform, we effectively established a lingua franca where business rules and logic could be understood by both business and IT teams. The most immediate benefit was a dramatic reduction in errors and translation time; we moved from three disconnected silos to two collaborating ones. This meant we could finally get a clearer, more controlled view of our decision-making logic. However, the process was still very manual. While the teams could understand each other, every decision still required a collaborative, human effort. The core challenge of agility and scaling remained because you couldn’t rapidly test dozens of scenarios or let the system optimize itself; you were still limited by the sheer time it took for people to convene, discuss, and implement changes.
Agentic AI is positioned as a layer on top of existing decision management tools. Can you walk through how a semi-autonomous agent translates a business leader’s goal, stated in plain language, into an optimized and actionable recommendation for the company’s systems?
This is where the real magic happens. Think of the agent as an incredibly smart, tireless collaborator. A business leader can simply state a goal like, “We need to improve our net promoter score without increasing our loss ratio in the Midwest region.” The agentic AI, which sits right on top of the decision management platform, takes that plain-language objective and gets to work. It autonomously assesses the existing data and rules, translates the goal into analytical components, and simulates thousands of potential adjustments to pricing, underwriting, or product features. It then presents an optimized recommendation, not as an abstract idea, but as something concrete that IT and data science teams can immediately understand and vet. The grunt work of assessment and translation is handled for you.
Imagine a CFO asks an AI agent, “What is the profitability impact of tightening underwriting criteria for coastal properties by 15%?” Explain the process the agent uses to run that simulation and how this immediate feedback loop fundamentally changes strategic financial planning for insurers.
This is a game-changer for financial strategy. When a CFO poses that question, they are no longer kicking off a multi-week project. Instead, the AI agent instantly taps into the decision management system. It runs that specific scenario—tightening the criteria by exactly 15%—across thousands, or even millions, of customer profiles in a simulated environment. Within moments, the CFO sees a clear dashboard showing the projected impact on premiums, the expected change in the loss ratio, and the overall effect on profitability. This creates an immediate feedback loop. Instead of making strategic decisions based on guesswork or historical data, leadership can now actively “war-game” scenarios, test hypotheses in real time, and instantly gauge the financial benefits before a single line of code is changed in the live system.
Some may worry that democratizing decision-making makes technical teams less critical. How does agentic AI actually elevate the roles of IT and data science, and what new types of high-value work can they pursue when freed from routine translation tasks?
It’s a common misconception, but the reality is quite the opposite. Agentic AI doesn’t replace these skilled professionals; it elevates them. For too long, we’ve used our brilliant data scientists and IT experts as simple translators, bogged down in the tedious work of converting ambiguous business requests into functional code. By automating that translation, we free them to focus on truly high-value work. IT can now concentrate on building a more robust, secure, and scalable architecture, ensuring governance guidelines are flawlessly integrated. Data scientists can focus on pioneering new predictive models and exploring deeper, more complex business challenges instead of just parsing dashboards for routine queries. It transforms them from interpreters into strategic architects of the company’s future.
For an insurer ready to adopt this technology, what is a practical, high-impact first use case to implement? Please outline the key business metrics they should track to measure success and justify further investment, drawing on results like improved retention or NPS.
The best way to start is with an incremental, high-impact project. I would recommend focusing on pricing adjustments or the launch of a new, targeted product. These are areas where the speed and simulation capabilities of agentic AI can deliver very clear and measurable results. To prove its value, you need to track core business metrics. Look at the reduction in time-to-market for pricing changes—did it go from months to weeks? Measure the impact on portfolio performance, such as profitability and loss ratios in the target segment. And critically, track the customer-facing results. Some early adopters are already reporting incredible outcomes, like a 14% increase in customer retention and a 48% higher Net Promoter Score compared to their peers. Those are the kinds of powerful numbers that justify scaling the investment across the organization.
What is your forecast for the insurance industry over the next five years as agentic AI adoption separates the leaders from the laggards?
My forecast is that the gap between the leaders and the laggards will widen dramatically. The insurance industry is on the cusp of a major transformation, and as PwC reports, over half of insurance executives believe AI will be the most significant technology reshaping the industry. Those who invest now in layering agentic AI onto their decision management strategies will operate with a level of speed and precision their competitors can’t match. They will be able to react to market shifts in days, not months, and their strategies will be directly and tightly linked to execution. The insurers who wait on the sidelines, clinging to their siloed, manual processes, will find themselves unable to compete on price, product, or customer experience. In five years, agentic AI won’t be a luxury; it will be the defining characteristic of a successful, modern insurer.
