We have now done fourteen seed investments in European InsurTech since 2020. The evaluation framework has evolved — some things we weighted heavily in the early funds have turned out to matter less; some signals we treated as secondary have proven to be the most predictive. What follows is our current framework, as honestly as we can describe it.
I want to be specific because I think vague investment criteria — "we look for exceptional teams addressing large markets" — are genuinely unhelpful to founders trying to understand whether they should spend time building a relationship with a given investor. We are not trying to cast a wide net. We are looking for a specific type of company in a specific sector. This is what that looks like from our side of the table.
Question One: Do You Know Where the Data Comes From?
The first question is about the founding team's model of their own product. Most InsurTech products require access to insurance data — claims data, policy data, customer behavioral data, telematics — to function. The team needs to have a clear and honest account of where that data comes from, who controls it, and what happens to their product if data access is renegotiated.
The failure mode we have seen several times is a founding team that has modeled their product on carrier data access as an assumption rather than as a negotiated reality. The product works in the prototype because the team used publicly available data or data from a single cooperative carrier relationship. Scaling requires data from multiple carriers. The team discovers that carriers view their data as a competitive asset and are not eager to provide it to a startup that might be serving their competitors simultaneously. The business model does not survive contact with the procurement process.
What we want to hear: a detailed account of where the training data comes from, what the contractual relationship is, whether the data access is exclusive or shared, and what the product looks like if one data partner relationship ends. Founders who can answer these questions at seed stage — before they have revenue — are the ones who have done the hard thinking.
Question Two: What Is Your Theory of Distribution?
Insurance is a distribution-intensive industry. Getting a product to the policyholder requires navigating a chain of brokers, MGAs, price comparison websites, or direct digital acquisition — and each channel has different economics, different compliance requirements, and different switching costs. The founding team's distribution theory is load-bearing in a way it is not for most SaaS businesses.
The wrong answer to this question is "we will sell through the major brokers." That is not a distribution theory; it is a distribution aspiration. Brokers face limited incentive to recommend a newer product over an established one if the commission differential is not material and the operational integration burden is real. "We will sell direct via digital acquisition" is more promising but requires a view on customer acquisition cost in insurance, which is typically higher than in most consumer categories because insurance buying intent is low-frequency.
The right answer is a specific account of the first 5–10 distribution relationships the team plans to build, why those counterparties have an incentive to distribute, what the unit economics look like in each channel, and what the team has already done to validate distribution assumptions before raising a seed round. We have a strong preference for teams where at least one founder has operated inside the distribution chain they are trying to work with — because the knowledge of how that chain actually functions is not available from the outside.
Question Three: Can You Explain the Regulatory Risk?
European insurance regulation is an operational constraint that a founding team must understand before product launch, not after. The specific requirements vary by product type, distribution method, and geography — but they share a common structure: you need to understand whether your product requires authorisation, what conduct standards apply to your sales process and claims handling, what capital requirements apply if you are taking risk, and what data protection requirements govern how you use policyholder data.
We are not looking for founders who have already solved all of these questions. We are looking for founders who have a clear map of the regulatory landscape their product operates in, who have identified the questions they do not yet have answers to, and who have a credible plan for getting those answers. A founding team that says "we think we need an IDD registration in three countries but we have not yet confirmed whether our specific product falls under the regulated activity definition" is ahead of a founding team that says "we checked with a lawyer and we are fine."
The first team has a specific, bounded question they are resolving. The second team has not asked the right questions yet.
What We Weight Less Than We Used To
Early in the fund, we weighted the technical sophistication of the founding team's ML models quite heavily. We have updated this view. Technical sophistication at seed stage is a necessary condition for the team to build a credible product, but it is not a differentiator in the way we initially believed. The ML tooling available to founding teams in 2022 is substantially better than it was in 2019. A competent team can build a technically respectable insurance ML product relatively quickly. The bottleneck is not technical sophistication; it is domain knowledge and distribution access.
We have also learned to weight less heavily the coherence of the founding team's narrative in the first meeting. Insurance is a domain where the most important knowledge is tacit — accumulated through years of operational experience — and tacit knowledge is hard to articulate cleanly in a pitch. Some of the most impressive founders we have backed were initially difficult to understand in the first conversation because they were reasoning from deep domain knowledge rather than a pitch narrative. The slow build — many conversations before we understood what they were actually building — has sometimes preceded the best investments.
A Note on Stage Discipline
We invest at seed. This means we are writing the first institutional check, or close to it. We are not looking for companies that have already proven their model at scale — those companies are at the wrong stage for us and should be talking to growth-stage investors. We are looking for companies where the founding team has done enough to convince us that the problem is real, the approach is technically sound, and the team has the domain depth to navigate the insurance distribution chain.
The seed stage in European InsurTech has specific characteristics that distinguish it from seed investing in other sectors. The sales cycles to carriers are long — 9–18 months from first conversation to live integration is normal. The regulatory clock starts ticking before revenue starts flowing. The initial product is often enterprise software, not a consumer product, which means revenue ramp is slow even with a successful pilot. All of this means the company needs more runway to reach meaningful proof points than in a typical consumer SaaS seed investment.
We account for this in how we structure our initial checks and our reserve capacity. Founders who know this from their own research — who have modeled a runway assumption that reflects the actual enterprise sales cycle in their specific distribution channel — give us confidence that they understand the sector they are building in.