Fund II closed in early 2024 with €54M in commitments. The two-year gap between Fund I and Fund II is enough time to have learned things that change how we invest — not to the point of abandoning the core thesis, but enough to recalibrate specific aspects of it. Writing this down is partly for our own discipline and partly because the founders we talk to deserve to understand what has changed in our thinking since the original framing, which was written in a different market environment with a different portfolio knowledge base.
The first thing that changed: our view on the go-to-market timeline for enterprise automation products. Fund I was written with an implicit assumption that a well-designed product with clear ROI would move through enterprise procurement in six to twelve months. The Fund I portfolio has taught us that twelve to eighteen months is more realistic for a first contract at a company above 250 employees, and that this timeline is relatively stable across product categories — it is driven by the buyer's internal governance dynamics more than by product quality or commercial intensity. This changes the capital requirements calculation: a seed-stage company targeting large enterprise needs either a longer runway before needing to show revenue scale, or a more disciplined approach to using a mid-market beachhead as the first growth engine before moving up-market. We now weight this explicitly in how we think about check size and milestone expectations.
The second change: more conviction on horizontal platform potential, and more skepticism about narrow vertical point solutions as terminal business cases. This is not a reversal — narrow point solutions can be excellent seed investments precisely because the product focus produces faster time-to-first-customer than a platform with broad ambition. But we have observed in the Fund I portfolio and in the broader market that the most capital-efficient path to significant ARR usually involves a single workflow surface that is deep enough to command substantial contract values, rather than multiple shallow workflow surfaces each commanding small contracts. We are now actively looking for initial scope that is narrow enough to be buildable at seed stage but sits within a workflow surface that could support €5-10M ARR without requiring a platform rebuild.
The third change is about the founding team profile. In Fund I, we were willing to back founding teams where the technical capability was very strong but the commercial experience was minimal, on the thesis that commercial skills can be hired and technical domain knowledge is harder to acquire. This is still true, but we have updated the threshold. The minimum we now look for is evidence that at least one person on the founding team has done a meaningful number of discovery conversations with potential customers before writing any code — not as a product management exercise, but as a genuine effort to understand whether the problem they are solving is urgent enough that a buyer would prioritise it over competing budget requests. The companies in our portfolio that struggled earliest were those where the founders built in isolation from potential buyers and discovered post-build that the commercial dynamics were different from what they had assumed.
What has not changed: the core thesis that AI-native workflow automation for European mid-market enterprise is the most consistently high-quality seed investment category we have access to, given our specific knowledge base. The market dynamics we described in 2021 — the gap between operational sophistication and software modernity in DACH mid-market companies, the structural advantages of European-origin products in data sovereignty, the network dynamics of German enterprise reference selling — all remain intact, and in several cases have become more pronounced as the supply of well-designed automation products has increased but the buyer base has remained relatively stable in its procurement culture. The fund II thesis is the Fund I thesis run at a higher confidence level, with adjustments to the team composition criteria and go-to-market timeline assumptions that better reflect what we have learned.