Mosaic raises $18M Series A to automate private-market financial modeling with AI

Mosaic, a platform that uses AI to automate financial modeling for private market deals, has raised an $18 million Series A led by Radical Ventures. The round includes strategic investor Troy Pospisil, CEO of Ontra.ai, and brings Radical Ventures partner Ryan Shannon and former Apax CEO John Megrue onto the board and advisory team. The company's system ingests deal documents like CIMs and generates LBO, DCF, and other models within minutes, replacing manual Excel work that can take hundreds of hours per deal. Funds will go toward expanding workflows for private equity and moving into investment banking and private credit.

Why it matters: Mosaic exemplifies the recurring pattern of vertical AI automation in financial services, where LLMs are applied to narrow, high-value workflows that have long been dominated by manual spreadsheet modeling. The company is not a foundation model provider but a layer-two application that combines deterministic computation with AI document parsing—a formula that has proven sticky in legal and compliance verticals (e.g., Ontra, Ironclad). Radical Ventures' involvement signals that AI-native financial infrastructure is becoming a thesis for specialist AI funds, not just mainstream VCs. The deal also updates the player map for AI in finance/operations, adding a dedicated private-market modeling tool to a segment that includes players like 9fin (fixed-income analytics) and Acompany (M&A advisory).

The expert take: Mosaic's approach—deterministic engine plus AI ingestion—sidesteps the hallucination risk that plagues pure-LLM financial analysis. By targeting auditability and reproducibility, it meets the compliance requirements of PE and investment banks, which are often slower to adopt generative AI. The $18M round size is modest in AI terms, but the strategic hires (Megrue from Apax, Pospisil from Ontra) suggest an emphasis on distribution and enterprise trust. If Mosaic can expand from PE into credit and IB without sacrificing model accuracy, it could become the de facto standard for deal modeling—a market currently fragmented across Excel templates and boutique providers. However, the vertical remains nascent, and competition from incumbents like Bloomberg or FactSet integrating AI features is a clear risk.