The insurance industry’s capacity to provide coverage for climate and cyber risks is increasingly the subject of public and regulatory attention. What receives less scrutiny is the structural reason the industry struggles to price and scale these exposures: fundamental gaps in the data needed to model them, combined with organizational boundaries that prevent the kind of information sharing that would make better modeling possible.
Key takeaways for insurance leaders
- The pricing prerequisite: Climate and cyber risk cannot be priced with confidence, and therefore cannot be covered at scale, until the industry closes the underlying data gaps that proprietary, siloed models were never built to fill.
- The AI reframe: AI’s most consequential use here is not generative applications or document processing. It’s aggregating disparate data sources into richer risk databases, giving modelers more to work with and faster cycles to stress-test against.
- The ReSource Pro view: Data governance, standardized reporting, and cross-functional information management are the internal discipline that lets an organization show up credibly to industry-level collaboration, not a back-office afterthought.
Dawn Miller, CEO of Lloyd’s Americas and Chief Commercial Officer of Lloyd’s, addresses this directly in a conversation with The Insurance Lead. Her argument is not abstract. It is a precise diagnosis of why problems at the scale of climate change and systemic cyber exposure require something different from how the insurance industry has historically operated.
Why scale demands collaboration that the industry has resisted
Traditional peak peril models were built and refined within individual organizations. They reflected proprietary assumptions, internal loss data, and competitive boundaries that made sharing inconvenient. For standard commercial perils within well-understood return periods, that approach has worked adequately. For climate risks that are accelerating outside historical parameters, and for cyber exposures that are correlated across entire industries and geographies, it is no longer sufficient.
Miller’s framing is direct: “The challenges ahead are too large for a siloed approach.” Capital cannot be deployed at the scale these risks require without the transparency that allows it to be priced with confidence. Data deficits prevent proper pricing of emerging exposures, and proper pricing is a prerequisite for coverage availability. This is not primarily a technology problem. It is a structural and cultural one. The industry has historically treated data as a competitive asset to be protected. At the scale of climate and cyber, that posture creates coverage gaps that are bad for policyholders and ultimately bad for the market.
Where AI actually creates value in this context
The conversation around AI in insurance modeling often gets captured by generative applications, chatbots, submission processing, and document handling. Miller identifies a more consequential application: aggregating disparate data sources into richer risk databases that support better assessment and modeling. The computational power AI enables in analytics is a significant leap beyond what actuarial teams could run on historical data infrastructure.
For CROs and heads of catastrophe modeling, this is a meaningful practical signal. The value of AI is not replacing the judgment of experienced modelers. It is expanding the data those modelers can work with and accelerating the analytical cycles that allow models to be stress-tested against a wider range of scenarios. Climate risk in particular involves compounding variables that traditional models were not designed to handle simultaneously. AI-assisted modeling does not solve the underlying data gap, but it does make better use of the data that is available while the industry works on sourcing more.
Proof that scaling works when the conditions are right
Miller cites Whisker Labs as a concrete example of an innovation that moved from solving a specific problem to deploying at scale across the United States. The significance of this example is not the specific technology. It is the proof point that insurance-adjacent innovations can achieve real reach when the organizational and capital conditions support them. Scalable deployment requires more than a good idea. It requires capital transparency, operational infrastructure, and the institutional willingness to move from demonstration to commitment.
A ReSource Pro perspective
Closing protection gaps in climate and cyber risk is not a problem that any single organization solves independently. It requires the kind of cross-functional data discipline and operational credibility that allows organizations to participate in industry-level collaboration efforts. Insurance organizations investing in data governance, standardized reporting, and cross-functional information management are building the internal conditions that make external collaboration possible. That foundation matters more as the complexity of the risks the industry is being asked to cover continues to increase.
Source: Innovation only works when minds come together: Lloyd’s Dawn Miller on breaking down silos
Featuring: Dawn Miller, CEO Lloyd’s Americas and Chief Commercial Officer, Lloyd’s
Publication: The Insurance Lead
Original Publication Date: March 17, 2026
