Velvet's Workflow Intelligence for Venture
Alex Johnson, co-founder & CEO of Velvet, on vertical AI for venture capital
This points to the real prize in venture software being proprietary market data, not memo writing. Aumni became valuable because it turned messy legal packets into a structured dataset JPMorgan could own. Velvet is trying to go one layer deeper, using AI to structure not just legal docs but decks, data rooms, CRM activity, and investor decisions into a live map of how deals get evaluated, won, and eventually traded.
-
Aumni was a document intelligence business. Velvet is aiming for workflow intelligence. That means capturing who looked at a deal, what questions they asked, what comps they pulled, who they introduced, and whether they invested. That behavioral data is much harder to replicate than a one time legal document parse.
-
The comparison also explains why Velvet is pairing software with secondaries. In private markets, the biggest bottleneck is still matching buyers and sellers in an opaque market run through brokers and spreadsheets. A daily use diligence product can become the system where both data and transaction intent first show up.
-
The closest analogue is Bloomberg in public markets, not a memo bot. Bloomberg wins because research, messaging, market data, and counterparties sit in one workflow. Private markets still split those jobs across Affinity, Carta, PitchBook, brokers, and ad hoc files, which leaves no single operating system for decision making or liquidity.
Where this heads is toward venture tools that do not just help analyze a company, but become the access layer for private capital itself. If Velvet can keep compounding first party deal and behavior data, it can move from helping funds decide faster to helping institutions, family offices, and RIAs find liquidity and allocation in the same workflow.