Why AI won’t ruin insurers’ secret sauce (and how it can scale it)

Behind closed doors, many insurance leaders are expressing the same questions about how AI might impact their competitive advantage: When competing insurers use similar AI solutions, do they all end up writing the same book of business, or pricing a deal similarly? Does increased AI adoption evaporate an insurer’s unique competitive moat? 

In reality, AI doesn’t replace an insurer's proprietary strategy, it amplifies whatever strategy is already there, for better or worse. 

Carrier's most valuable asset is buried

Most carriers have spent years, sometimes decades, perfecting their unique approach to underwriting risk. That “secret sauce” is their product definition and strategy: appetite, guidelines, judgment about which risks to take and how to price them. This strategy doesn't often live in one place. Instead, it's split between legacy policy administration systems (PAS), where years of guidelines got hardcoded, and a patchwork of other tools and manual processes that disconnected from each other. 

Some of the most valuable knowledge also lives in the heads of tenured underwriters, who know from experience that ‘coverage A’ paired with ‘exclusion B’ is a recipe for strong profitability. Because this knowledge often goes undocumented, it walks out the door when they retire.

Extracting all of this institutional knowledge is like an archeological dig. When information about strategy is scattered across an old core system, a dozen disconnected tools, and people’s minds, it becomes difficult to inspect, evolve, or apply consistently. 

Adding AI on top of that fragmentation doesn't fix the problem. It magnifies those conditions and makes your operation consistently inconsistent at a blistering pace. 

In her session at the Federato Insurance Leadership Forum, Cassie Kozyrkov, former Chief Decision Scientist at Google and CEO of Kozyr, diagnosed this problem as "random acts of AI": when organizations deploy AI tools without a coherent decision-making framework underneath them. According to Kozyrkov, AI amplifies decisions, but doesn’t improve them. An organization with poor decision infrastructure will make poor decisions faster, and at greater scale, when it adds AI to its stack. 

That's the warning. Now here’s the opportunity. 

Leveraging an AI-native platform as a strategic multiplier 

According to BCG’s research, AI spend in P&C is tripling, but only 38% of insurers are seeing ROI. The ones who are pulling ahead have figured out that AI is a strategy multiplier, not a replacement for strategy.

Two carriers running the same platform on the same submission will produce different outputs. The AI use may be the same, but their guidelines, their portfolio context, and their underwriting judgment are fundamentally different. That's the secret sauce — and AI can scale it. But only if they’ve done the work of making that judgment explicit.

This is where the carriers getting the most value from AI are separating from those who aren't.  Instead of adopting random acts of AI on top of buried or hardcoded strategy, the winners are replacing their rigid PAS systems with an AI-native platform that structures their strategy into a data foundation AI agents can use consistently at scale. 

How Federato makes buried strategy actionable

Federato's AI-native platform is designed to do exactly that: turn strategy that's currently buried into strategy that's structured, easy for leadership and product managers to modify, and usable by AI. The platform does this in two ways. First, it extracts information from source documents and systems and puts it into a structured data foundation with a consistent schema. Then Product Studio and Control Tower expose that configuration in a way that's easy to see and change, so it doesn’t stay buried or hardcoded. 

Federato’s Product Studio establishes the foundation for underwriting discipline and dictates what the business can write. It gives product teams a single, central place to manage rates, coverages, policy forms, and eligibility. Because every policy in-force is automatically linked back to the exact rules in place when it was bound, carriers gain a built-in audit trail and stronger regulatory defensibility, reducing compliance risk without added manual work.

Control Tower gives underwriting leaders real-time control over what the business should write, right now. Market conditions shift constantly, and a risk that fits the product on paper may not fit the portfolio's current goals. With Control Tower, leaders can adjust appetite instantly in response to those shifts. For example, if wildfire exposure spikes in California, leadership can immediately tighten guardrails, routing high-hazard risk to review while steering the business toward opportunities that improve geographic diversification.

Because agentic underwriting draws on these updated priorities in real time, frontline execution stays aligned with leadership's strategy at all times, not just at the next planning cycle. This matters because guideline changes today often arrive piecemeal, through emails, uploads, or memos, leaving some underwriters working from outdated guidance while others have updated information. When the AI agent draws from a single, real-time source of truth, that inconsistency becomes far easier to manage and correct.

Together, the two solutions give carriers both stability and agility: a defensible product foundation that protects against regulatory and compliance risk, and a responsive control layer that lets leadership actively shape portfolio outcomes as conditions change, rather than discovering drift after the fact.

This shared foundation provides carriers with new visibility into when and where underwriters deviate from AI recommendations, revealing whether strategy needs updating or teams need alignment. 

Institutional knowledge with no retirement date 

The carriers who fall behind will be the ones who let the recipe for their secret sauce remain implicit, locked in the heads of senior underwriters who will eventually retire, and in the depths of legacy code nobody can read.

The winners will be the ones who made that knowledge explicit, structured, and scalable. Once it's encoded well enough for AI to implement consistently, it doesn't leave your organization when the senior talent does.

AI doesn't write the same book for every carrier who adopts it. It writes the book you've defined. Carriers who use an infrastructure that matches line-of-business strategy to front-line execution can drive premium volume up while keeping their combined ratio locked down.