What Northbarn Looked for in Fund II: How Our Thesis Evolved from 2019
When we raised Fund II in late 2022, closing at $115 million in November, our thesis had evolved in three important ways from when we raised Fund I in early 2019. The regulatory environment had shifted, the quality and accessibility of AI tooling had changed materially, and we had four years of portfolio learning to draw on. We explain those evolutions here.
The first evolution: generative AI changed the product surface for compliance-native companies. When we raised Fund I, the AI capabilities most relevant to regulated-industry workflows were narrow: classification, extraction, and prediction from structured data. These were valuable but limited. By the time we were evaluating Fund II investments, large language model capabilities had dramatically expanded what AI could do in clinical documentation, legal research, and financial analysis. The compliance architecture question did not change — companies still needed to build products that were auditable, defensible, and institutionally trusted. But the product surface that compliance architecture could be applied to had grown enormously.
The second evolution: the Series A market in regulated-industry AI became more competitive. In 2019, we were the only fund in Boston with an explicit thesis around healthcare and legal AI at Series A. By 2022, the category had attracted significant capital and the number of qualified Series A companies in our space had grown. We responded by sharpening our differentiation criteria: we now look more explicitly for founding teams with prior institutional relationships at the customer level, not just domain expertise. A healthcare AI company whose founders have existing relationships with clinical informatics leaders at target health systems is in a structurally different position than one that does not.
The third evolution: we became more selective about the healthcare sub-sectors where we believe AI creates durable advantage. Early in Fund I, we backed companies in clinical documentation, diagnostics, and patient engagement across a range of healthcare settings. In Fund II, we are concentrating in areas where the compliance architecture is the explicit source of product differentiation: revenue cycle, prior authorization, and specialty-specific clinical decision support. These are the healthcare workflows where the AI is not just improving efficiency but is creating verifiable compliance outputs that institutional buyers will pay significant premiums to obtain.