Somewhere, a claims file was never opened.
No loss figure is attached. No adjuster’s photographs. No restoration invoice. No family displaced for six weeks. No business interruption report.
The day passes without incident.
A warning arrived in time, someone acted, and what might have become a claim remained an ordinary day.
The next underwriting advantage will come from proving which mitigation actions change expected loss, rather than relying solely on accurately predicted loss.
Creating more days like this starts with understanding where losses occur and which actions could change the outcome. Rising claims costs make this urgent.
In the United States, ISO data published by the Insurance Information Institute shows that the average homeowners claim rose from $13,884 in 2019 to $20,062 in 2023. In the UK, the average home insurance claim rose 15% year over year to £6,000 in 2025, according to the Association of British Insurers.
Water damage shows the cost of risks where earlier action could help. US water damage and freezing claims averaged $15,400 over 2019–2023, according to the same ISO dataset. Separately, LexisNexis Risk Solutions reports a 63% rise in non-weather-related water claim severity between 2019 and 2025. In the UK, escape-of-water claim severity rose 14% in 2024, with average costs reaching a record £5,800 in Q1 2025, according to LCP’s analysis of ABI data.
Prevention should become an underwriting advantage that expands the risks insurers can confidently cover. The opportunity for insurers is to connect risk signals, timely action, and outcome evidence across their existing workflows then test whether those interventions actually change loss.
From risk insight to intervention
Start with a preventable loss, a clear mitigation action, and a way to measure whether the benefit exceeds the cost. Forecasts and connected devices help identify when that action could matter.
Consider frozen pipes. Property characteristics and temperature forecasts can help identify buildings that need attention, while connected monitoring can detect temperature drops. A timely alert could prompt a homeowner to maintain heating or a facilities team to inspect a vacant site.
Now consider a commercial warehouse facing a flood warning. A location-specific forecast could trigger a response plan to deploy barriers at vulnerable entrances and move stock out of harm’s way, before assigning a facilities manager to confirm the work. Even if water enters the building, those actions could reduce damage and shorten disruption.
A 2021 study published in the Bulletin of the American Meteorological Society helps explain why the response matters. Analysing 4,468 household loss cases across six German floods between 2002 and 2013, researchers used matched comparisons to estimate an average reduction of approximately €3,800 in contents damage when households received at least an hour’s warning and knew what to do. Warnings alone did not produce a statistically significant reduction.
For the warehouse, the plan must account for existing protection, warning lead time, and who can act safely. Someone needs responsibility and the means to respond. The insurer then needs evidence that the action reduced loss.
Measuring what did not happen
Connecting a risk signal to a specific mitigation action and a claims outcome builds a credible business case for prevention at scale. A year without a claim is insufficient evidence on its own.
For the warehouse, that means recording site flood conditions, warning delivery times, mitigations taken, and subsequent damage or interruption. A site the flood never reached tells us little about whether its barriers worked. Records of exposure and action matter even when no claim is made.
Across a portfolio, insurers can compare sites with similar exposure using matched comparison groups or phased rollouts. Models of claim frequency and severity should account for flood intensity, property characteristics, exposure periods, and differences between customers who participate and those who do not. Accurately predicting a loss does not establish that an intervention prevented it.
The analysis should distinguish between an unreceived alert, an unexecuted action, and an ineffective safeguard. Each points to a different operating decision, from improving delivery to revising the intervention itself. Comparing estimated loss reductions with monitoring and service costs can clarify whether to scale a program or refine its design.
The difficulty of establishing that impact has commercial consequences. EIOPA’s 2023 impact underwriting report, based on a pilot involving 31 European insurers, found that only a few explicitly incentivized climate adaptation through premium discounts. Difficulty assessing the effect on risk was among the reasons. Individual avoided claims remain uncertain, but program-level effects can be estimated and tested.
Connecting prevention with underwriting and pricing
Ignoring verified mitigation can overprice well-protected risks, while rewarding untested measures can underprice the exposure. Credible evidence of reduced loss can change both an insurer’s willingness to cover a risk and the price it charges.
The warehouse’s flood exposure may initially put it outside an insurer’s appetite. Combining verified barriers with a tested response plan could make it acceptable on defined terms. That assessment must also account for the insurer’s total exposure to the same flood event across its portfolio.
Underwriters would need current evidence that the safeguards suit the site and remain operational, with clear responsibility for verification. They would use that evidence when deciding coverage and deductibles. Pricing teams would assess the remaining expected loss and service costs before considering a premium adjustment, subject to applicable rating requirements.
That evidence could also shape the product. A policy could combine flood cover with monitoring and an agreed response service. Its terms should explain responsibilities for maintenance and action, how equipment failures or unanswered alerts are handled, and what evidence is reviewed at renewal. Customers should understand how participation and safeguards affect their cover and pricing, and how their data will be used.
Participation cannot be assumed. Guidewire’s 2026 European Insurance Consumer Survey, covering the UK, France, Germany and Spain, found that 34% of respondents understood why insurers wanted data from sensors and connected devices but would prefer that it was not collected. Insurers need to make the benefit of sharing that data clear.
For customers, the value is practical support that helps reduce disruption and sustain access to suitable cover. Insurers should consider assistance for policyholders who cannot install equipment or respond themselves, while keeping warnings relevant enough to sustain participation. For mutual insurers, this also offers a tangible way to support members before they need to claim.
The technology foundation prevention needs
Turning prevention evidence into better insurance decisions requires a modern platform that connects core processing, business context, and intelligence. Guidewire provides that foundation. For prevention, evidence of exposure and mitigation efforts should remain connected to the insured risk throughout the policy lifecycle, even when no claim occurs.
Guidewire UnderwritingCenter uses purpose-built AI agents to surface relevant information, helping underwriters apply their expertise to complex commercial risks. Alongside that assessment, Guidewire PricingCenter enables pricing and actuarial teams to build, test, and deploy advanced models that can incorporate credible evidence of mitigation.
PricingCenter’s native connection to Guidewire PolicyCenter on Guidewire Cloud brings approved rates into quoting and renewal. Guidewire ClaimCenter supports the full claims lifecycle, capturing claims activity and loss information to help insurers evaluate outcomes. Guidewire Data Platform makes this core data available for analysis, supporting learning across the policy and claims lifecycle.
Prevention also relies on connecting this history with new risk signals and customer action. Guidewire Marketplace provides access to validated integrations for specialist capabilities, while Guidewire AI brings intelligence into workflows. For the warehouse, appropriate integrations and workflows could retain the flood warning, site exposure, and confirmed response against the policy record, whether or not a claim follows. That evidence could inform renewal decisions and analysis across similar risks.
From prevention to resilience
For the warehouse, resilience could mean less stock destroyed and a shorter interruption after flooding. For a household, it could mean avoiding a burst pipe and remaining at home through a freeze. The next step for insurers is to identify where those outcomes are achievable and how to make them repeatable.
The Future of Risk experience at Connections 2026 gives insurers an opportunity to meet companies working on those challenges and examine the interventions they enable. The roster includes LeakBot, MākuSafe, Previsico, Toborlife, and spacebands, with approaches spanning water leaks, flood, and workplace safety.
Meet the featured companies with a decision from your own portfolio in mind. Underwriting and pricing leaders can explore which controls could change risk appetite or coverage terms. Claims leaders can ask how interventions and outcomes are recorded, including when no claim is made. Risk leaders can examine who responds to a warning and how the effectiveness of that response is verified.
In Part Two, we will look more closely at the featured companies, the risks they address, and how their interventions could shape underwriting, pricing, and product design.
For the customer, success may look like a day when nothing happened. For the insurer, the opportunity lies in establishing which interventions reduce loss and applying that evidence to improve underwriting, pricing, product design, and customer experience. This transforms prevention into an underwriting advantage, expanding the risks insurers can confidently cover.