Small business insurance has been the awkward middle child of the insurance market for thirty years. Too large and varied for personal lines automation. Too small and fragmented for the bespoke treatment given to commercial mid-market. Brokers tolerated it. Insurers priced it poorly. Founders tried and failed to digitize it. Until recently, the timing was never right — and the reasons why are worth understanding precisely, because they tell you why now is different.
Why the Decade Delay Was Structural, Not Lazy
The personal lines digitization story — direct comparison sites, telematics, instant-issue motor and home policies — followed a particular logic: relatively homogeneous risk, large volumes of historical loss data, regulatory frameworks designed for standard policies, and consumers who broadly understood what they were buying. Aggregate the data, build a pricing model, create a clean digital journey, reduce distribution cost. The playbook was clear.
SME insurance breaks most of those conditions simultaneously. A 12-person software consultancy, a 50-seat restaurant, and a three-van plumbing business all fall into "small business" brackets but carry entirely different liability profiles, occupancy risks, and claims patterns. Aggregating them into a single digital pricing model historically produced either wildly inaccurate risk selection or products so broadly written that margins collapsed whenever claims arrived from the tail segments you had not adequately modelled.
The data problem compounded this. Commercial property surveys and occupancy classifications were maintained in formats that pre-dated relational databases. Broker systems — BSQUARE, Acturis in the UK, various Scandinavian incumbent platforms — stored SME risk data in ways that made ML feature extraction genuinely difficult. There was no shortage of historical claims data in the aggregate; the challenge was matching it to the right categorical descriptors at the individual risk level.
Three Conditions That Have Now Changed
What shifted is not a single technological breakthrough. It is a convergence of three structural conditions that each needed to reach a threshold before the SME digitization thesis became investable.
Open banking and business data APIs. The PSD2 directive, live across EU markets since 2019, created the infrastructure for authorised access to business transaction data. An SME insurer can now — with appropriate consent — assess actual revenue patterns, payroll exposure, and seasonal trading cycles from bank feeds rather than relying on self-declared turnover figures at application. This does not eliminate information asymmetry, but it materially improves it. For SME liability and business interruption pricing, actual revenue data is substantially more predictive than declared revenue categories. We saw this first deployed at scale in Sweden and the Netherlands before Copenhagen-based carriers started building towards it.
Companies House and business registry data quality. Across most European markets, the quality and accessibility of official business registration data — sector codes, director relationships, registered addresses, filing history — improved significantly between 2018 and 2022. Cross-referencing a business application against registry data catches mis-stated sectors and identifies connected-party risk that manual broker screening missed. For fraud containment at application stage, this matters considerably.
The SME digitisation of the SMEs themselves. The pandemic forced business owners who had never used digital administration tools to adopt cloud accounting, digital payroll, e-commerce platforms. The average SME in Denmark, the Netherlands, or Ireland now has substantially more structured digital data footprint than in 2018 — and the accounting integrations to share it. An insurer offering a premium adjustment tied to actual revenue rather than projected revenue, reconciled through Xero or Dinero at year-end, has a genuine product proposition that did not exist five years ago.
Where the Product Innovation Is Actually Happening
When we look at where the interesting companies are emerging — including work we saw from the Superscript team when we invested in 2021 — the product architecture is not simply "personal lines but with a business name on it." The genuine innovation is in bundling and dynamic packaging.
A professional-services SME buying cyber liability, professional indemnity, and employer's liability as three separate policies from three separate carriers, renewed on three different anniversary dates, managed through a broker who charges 20-25% of premium, is an obviously inefficient structure. The opportunity is building the platform layer that holds the carrier relationships and MGA licences, then presenting the SME with a single point of purchase, a single premium, and a single renewal. Superscript's model approaches this. Qover approaches a version of it from the embedded distribution side. The infrastructure is still being built, but the direction is clear.
We are not saying that brokers are simply being replaced — the complex end of SME commercial, where a manufacturing business with product liability exposure needs genuine advisory input, is not going digital overnight. The segment that is moving is the under-50-employee, service-sector, lower-complexity end of the SME market, which represents a substantial portion of total SME premium volume by count.
The Regulatory Moat Is Real
One pattern we observe in founders who succeed in SME digital insurance versus those who underestimate it: the successful ones treat European regulatory complexity as a product feature rather than a compliance burden.
An SME digital insurer operating across multiple European markets needs carrier arrangements — or MGA status — in each jurisdiction, products approved under local supervisory requirements, GDPR-compliant data handling for business owner personal data, and Solvency II-consistent underwriting documentation if they are bearing any risk themselves. Founders who build the MGA structure early, who hire legal and compliance from insurance backgrounds rather than general fintech, who engage supervisory authorities proactively — these teams have a structural advantage that takes two to three years to replicate. The regulation is not a moat if you ignore it; it is a moat precisely because it is hard to replicate at speed.
What the Loss Data Will Show
The cohort of SME digital insurers that launched between 2019 and 2022 will be entering the phase where meaningful claims history starts to accumulate at a volume large enough for actuarial review. This is the test that personal-lines-trained digital insurance teams have historically failed — the SME segment they wrote was selected worse than they modelled, and loss ratios in years three to five exceeded projections.
The teams that survive this test will be the ones who built pricing on the structural data advantages described above — transaction data, registry data, real business digital footprint — rather than on a slicker application journey and cheaper broker-bypassing distribution. The distribution saving is real, but it does not protect a loss ratio if the risk selection was flawed from the start. We watch the emerging loss experience on 2020-2022 cohorts carefully; it will tell us more about which founders understood insurance versus which ones understood tech-enabled distribution.
The SME moment is genuinely here. The question that separates good investments from interesting products is whether the founding team is thinking about it from the underwriting model inward, or from the interface outward.