deep, operator-led access across clinical environments, hospital operators, health plans, and the investors who write follow-on rounds, and we apply that same lens to the companies we back, looking for channel ownership, workflow embedment, regulated relationships, and proprietary data loops that compound from real users. This thesis matters now because models are commoditizing; as frontier capabilities collapse toward parity and the marginal cost of intelligence approaches zero, the winners will not be the teams with the best model but the teams with the deepest distribution wedge, since technical differentiation increasingly buys a six-month lead while distribution moats compound for decades. Our conviction concentrates in two domains, applied AI translating model capability into measurable workflow value in regulated, high-friction enterprise environments, and healthcare across clinical workflow software, payer-side automation, care delivery, regulated diagnostics, and the operational backbone of US health systems. We invest pre-seed through Series A, partnering with founders before a clear category leader has emerged, and we are most useful to teams who think about reimbursement, contracting, regulatory positioning, channel design, and embedded data loops with the same rigor they bring to product. We pass on horizontal AI plays whose only edge is model quality, on healthcare companies without a credible path through procurement and reimbursement, and on any moat that depends on staying ahead of the next model release; we invest where the moat survives the next ten.