For complex commercial risks, an underwriter can know the business, location and requested coverage and still lack the physical evidence needed to understand what could actually go wrong. Factories, warehouses and construction sites may require engineering surveys covering fire protection, machinery, operational controls and other hazards before an insurer can properly evaluate the risk. The evidence eventually reaches underwriting, but collecting it, interpreting it and turning field observations into usable reports can become its own bottleneck.
Nettle is raising capital around that gap. The London- and New York-based company has announced an oversubscribed $4.8 million Seed round led by MTech, with Project A Ventures following on with a super pro rata investment and participation from Sure Valley Ventures, Portfolio Ventures, Ventures Together and selected angels. Nettle says the round brings total funding to $6.8 million and will support expansion across the US and Europe alongside growth of its engineering and go-to-market teams.
Founded in 2024 by former QuantumBlack colleagues Jack Miller and Katya Kinane, Nettle is building what it calls an AI Workspace for Loss Control. The problem is more specific than automating insurance paperwork. Loss-control and risk-engineering teams collect physical evidence that can influence whether an insurer accepts a commercial risk, how it prices the policy and what improvements it asks a policyholder to make. Miller captures the operating logic in one line: “insurers can only act on the risks they understand.”
Nettle tries to expand how much of that risk an insurer can understand without turning every location into a full manual survey. Before a visit, its platform can combine insurer data with external information such as natural-catastrophe data, standards and previous loss-control material. In the field, teams can capture photographs, documents and audio notes through guided inspections. The system then structures that evidence into risk findings, scores, recommendations and underwriting briefs, while keeping human review in the workflow. Nettle also supports third-party survey data, guided self-inspections and collaboration across more than 35 languages.
That architecture also explains why Nettle has spent time integrating external hazard intelligence rather than treating AI as the only source of risk data. Partnerships announced with 7Analytics and alkazar brought high-resolution flood and natural-catastrophe information into the platform, allowing physical evidence and external hazard data to sit inside the same assessment workflow.
The workforce argument behind the product is significant, although some of Nettle’s headline statistics remain company-sourced. Its funding announcement says risk-engineering teams can face inspection backlogs of up to six months and estimates that roughly 40% of risk engineers could retire by 2030. Separately, a Nettle report written by veteran risk engineer Peter Eymael modeled approximately $214,000 in additional costs beyond salary to develop a risk engineer over two years. Those figures help explain the company’s emphasis on preserving specialist knowledge, but they should not be presented as independently established industry benchmarks.
The more useful evidence comes from deployment. Nettle markets productivity improvements of up to 5x, but its named Allianz Türkiye case provides a more concrete boundary: during a pilot, the insurer’s risk-engineering team completed property inspections 2–3 times faster, with structured outputs reaching underwriters on the same day. Allianz Türkiye subsequently moved to a full enterprise deployment. That distinction matters because a named customer result is stronger evidence than a platform-wide performance claim.
Brotherhood Mutual is another named customer. Its risk-control leadership has framed the value less around replacing engineers than shifting their time away from report writing and toward policyholders. That is increasingly what Nettle’s product design suggests: use specialists for the judgment-heavy risks, while software handles more of the preparation, evidence organization, reporting and follow-up around them.
The next expansion pushes that model further. Nettle says the platform has moved beyond Commercial Property into Liability, Construction and Workers’ Compensation, while guided inspections can increasingly be completed by agents and policyholders for sites that may not justify sending a specialist. For insurers, the opportunity is greater portfolio visibility. The trade-off is that consistency, evidence quality and human oversight become more important as data collection moves beyond trained risk engineers.
That makes the $4.8 million round a test of more than geographic growth. Nettle has already shown that AI can compress parts of a commercial-property inspection workflow. The harder evidence now will come from whether those results remain repeatable as the platform moves across more insurance lines, more countries and more people collecting field evidence. If that works, loss control begins to look less like a scarce specialist service attached to individual surveys and more like a reusable intelligence layer running across an insurer’s portfolio.
