Private capital has long defied easy automation. It contains multiple asset classes (buyout, credit, venture, etc), based on unstructured data – much of the decision-making by GPs lives in the heads of the partners.
Solutions to automate how private capital works have typically begun with either the operations and accounting functions, or at the front office, in the investment process. Or they focus on a single niche asset type. Whatever the starting point, linking up the rest of the lifecycle or the market is difficult.
A new breed of startups is using agentic A.I. to be all things to all people, and create a data layer that serves as a record for the lifecycle across products. That’s the goal. Does the hype around agentic A.I. match the results?
To learn more, I spoke with Katriona Lee, founder of Reuben A.I., based in Sydney. Kat has a career in making private deals happen, either from the accounting and tech side (at Xero and Thompson Reuters), or as a banker (Barclays), or a consultant (EY, Accenture).
Kat spoke about why and how she’s founded Reuben A.I., the opportunity for GPs, and tricky questions around provenance, reliability, and the business model.
I should also confess, I forgot to include a question about the name of the company. In a pre-recording chat, Kat did tell me that story, and yes it’s about the Reuben sandwich…so apologies for my not including that in our interview, but if you’re interested in learning more, you can ask Kat yourself.
Timecodes:
0:00 - Jame’s intro to the issues around workflows in private capital
3:11 - Kat’s intro - opportunity, volatility, GP need to integrate context into systems
5:43 - The unstructured nature of private capital
7:39 - Becoming a founder to address problems of workflows and data
10:52 - Jame’s potted history of technology solutions and vendors in private capital
14:01 - What agentic A.I. makes possible across the value chain
17:33 - People versus agents within Reuben A.I. the company
20:01 - How to ensure repeatability and reliability when working with LLMs
23:24 - Who owns the data?
25:29 - How does provenance work in the context of A.I.?
28:29 - Beyond optimizing workflows to creating new types of insights
29:41 - Reuben A.I.’s founding, clients, funding, and costs of an LLM-dependent business
