Every party in an asset-based finance deal runs a system built for itself.
The originator’s was never built for the lender. The bank sees only the payment accounts.
So no two parties read the same numbers, and the inefficiency that follows is really a trust problem.
Recently, I sat down with Altin Kadareja, Co-Founder and CEO of Cardo AI.
Altin was born in Tirana, studied in Milan and spent 22 years in Italy before moving to New York last year. Insurance first, then commercial banks, then consulting, then BlackRock.
At BlackRock, he sat on the team created after the financial crisis to value what nobody could price: the loan pools and non-performing loans European banks were selling, and that U.S. funds were buying. Where the value sat, where the risk sat, what could be recovered.
All of it ran on spreadsheets.
Vendors kept pitching him software. But each one had the same answer: two to three months to change one column, six more to add an asset class. So he stayed in spreadsheets.
Eight years ago, he decided to leave and build it himself.
Cardo AI builds the layer that connects the originator, the lender and the bank, so every party reads the same numbers. Today, it runs inside some of the biggest names in private credit.
I asked him for a practical example.
One of his largest U.S. clients came to him after the First Brands collapse. It liked the credit quality of its receivables originator and couldn't see the fraud risk: 1,500 new invoices a week on a $50 to $60 million line, reconciled by hand against the invoice PDFs, the originator's spreadsheet and the bank statement.
In this episode of the Modern Capital Podcast, Altin and I cover:
Why only 30% of the $40 trillion ABF market is new collateral, and where the real innovation sits
Why banks underwrite on compliance and asset managers underwrite on return, and what that does to the loss curve
Two months at a bank against 20 days with an asset manager: why borrowers pay a premium for private credit
GPU financing, data centers and robotics: collateral the standard cash flow engines cannot model yet
Why a chatbot is 2% of what AI can do for an asset manager, and what has to come first
“I want to create a technology highway, which is software, data and AI, that everyone could work better in a standardized way, and could make the market scale faster.”
Altin thinks the market cannot scale faster until every party reads the same numbers. The spreadsheet is what stops them.
This is a conversation on how to fix it.
Altin’s case is that the foundation comes before the agents. Clean data, better workflows, then the agents that need both.
That sequence is one of the questions at The 2026 Private Markets AI Summit on November 4 in New York, where GPs, LPs and technology leaders building on agentic workflows, portfolio-level AI deployment and enterprise integration convene.