Cardo AI: End-to-end data management for private debt investment

29 Aug 2022

Altin Kadareja. Image: Cardo AI

Milan start-up Cardo AI is bringing investment management out of Word docs and Excel spreadsheets and into the world of data-driven decision-making.

Before he became the co-founder and CEO of Cardo AI, Altin Kadareja worked in several investment and risk management roles, including at New York-headquartered investment management firm BlackRock.

“Across these roles, I was very surprised to discover that billions of euros in private debt markets are evaluated, exchanged and monitored using only Excel files, Word documents or PDFs,” said Kadareja. He saw this as a surprisingly inefficient way for forward-thinking investors to operate and said it hampered decision-making.

“From the data collection to usage, modelling and the decision-making process, a lot of time was spent on Excel errors, late processes and exchanging files of different formats to make decisions,” he recalled. “It is impossible to scale in this way and dedicate such high costs to coping with legacy workflows.”

From that, Cardo AI was born. Founded in Milan by Kadareja and credit risk analyst Daniele Forza in 2018, this fintech start-up is “transforming the private debt market”, according to its CEO.

‘Excel is not enough anymore’

Kadareja said that while the private debt market has trillions of dollars of assets under management, it “is not standard, not liquid, and very fragmented, which makes it very complex to operate and continuously succeed in”.

“With increased competition and complex operational structures, market actors need to operate with speed and precision,” he said. “Excel is not enough anymore and hiring more risk managers and operational staff will not provide the elements to operate at scale.”

On the other side of the coin, Cardo AI’s challenge is ensuring potential clients see its services as the solution to these problems. Like all start-ups, this was an early hurdle, though it was helped over the hump by an unexpected aide. “It was difficult to attract customers due to longer education processes for prospective clients,” said Kadareja. “But the pandemic has accelerated digital adaptation and has reduced the time to market for new clients. Currently, we are stabilising our customer base in the EU and UK and in 2023 we will seek to expand our market in US and Asia.”

Cardo AI starts by aggregating and managing different sources of data, connecting to more than 90 external systems. It collects and cleans data from a variety of sources such as banks, payments providers, insurance companies, credit rating agencies, product reviews, social media and more.

“Then, these data are standardised,” said Kadareja. “This enables all users (from front-office to risk management, operations and investment teams) to easily interact, use and compare information against standardised investment performance metrics.”

The secret sauce, though, is made from Cardo AI’s proprietary machine learning models, which Kadareja said are attuned to the new world of fintech.

“We have witnessed the explosive growth of novel credit products like buy now pay later, revenue and inventory financing, salary financing, and so on. These products can’t run on old metrics regarding the probability of borrower default on a one-year period. To properly assess the risk of these products, new logics are needed.

“Models must consider the features and structure of the product: its loan characteristics, purpose, maturity, the structure of the underlying purpose on product’s usage, channels that the loan has been originated in, and several other elements combined. These can generate better predictors for borrower behaviour and the likelihood that they will pay back the loan in time. That is why, at Cardo AI, we have come up with ideas such as delay prediction models, propensity to pay back models and revenue limit estimations, etc.”

‘There is a talent war with big tech companies and other scale-ups’

Kadareja aims to keep innovating so that Cardo AI stays current, launching at least one major new feature or product every six months. Currently in development is a product focused on environmental, social and governance (ESG) criteria.

This new product will help Cardo AI clients to measure and analyse the ESG factors of their investments in detail. Kadareja also hopes to see more data sources added such as satellite data.

Another destination on the roadmap will see Cardo AI extend its offering to the private equity market. “In this way, we would offer a complete suite of technology and algorithms that enable our clients to structure and manage any debt or equity, to any borrower, from an individual to a small-to-medium enterprise or corporation, all at the click of a button,” said Kadareja.

All of this is powered by a team of more than 100 software engineers, data scientists, data engineers and business analysts. And getting to this point has not been easy.

“Team creation has been a big challenge,” said Kadareja. “There is a talent war with big tech companies and other scale-ups. It’s a fierce competition.”

In order to hold on to the team he has built, Kadareja has relied on keeping them connected. “The ‘great resignation’ and the pandemic turmoil made it difficult to build proper relationships with our team, but as soon as we had the opportunity, we organised meet-ups. Now, we have team retreats every six months, an in-person quarterly business review every three months, and more.

“The key is to listen to your talent and build trust. When this happens, you can discuss and carefully manage every need your team has,” he added.

10 things you need to know direct to your inbox every weekday. Sign up for the Daily Brief, Silicon Republic’s digest of essential sci-tech news.

Elaine Burke is the host of For Tech’s Sake, a co-production from Silicon Republic and The HeadStuff Podcast Network. She was previously the editor of Silicon Republic.