Inside the AI-powered VC with just three employees

Kadan Capital, the AI and fintech-focused VC firm, is probably as lean as VC firms go.

The two-year-old company has just three full-time employees, which includes former Antler investment manager Felix Frenzel, who is Kadan’s founding partner. Despite this, the firm has made eight investments so far this year, another six in 2025, and four the year before.

(From left) Kadan Capital founding partners Rei Murakami and Felix Frenzel alongside principal David Hwan / Photo credit: Kadan Capital

How does it manage such a pipeline? With a combination of automation and AI. In other words, it is applying AI the same way that it expects its portfolio companies to.

“As a VC, often I see people are still very analogous, very traditional, and it doesn’t make sense to me,” says Frenzel, referring to other investors.

Singapore-based Kadan believes an “AI-native” approach will help it stand out as Southeast Asia’s VC industry reaches a crossroads, with many investors still lacking the deep technical expertise needed to underwrite companies in the booming field.

As more firms pivot toward backing only “AI-fluent founders,” Kadan has gone a step further by building an AI and machine learning-powered investment engine modelled on the increasingly data-driven playbook of its US counterparts.

Weeding out the noise

According to Frenzel, Kadan seeks to invest in companies that “couldn’t have been built two or three years ago because the technology wasn’t there.”

One example is Lance, which uses computer-use agents to navigate hotel reservation systems as well as handle guest calls and internal messaging without requiring back-end integrations. These agents can “see” and interact with interfaces like websites and enterprise software, even when they don’t support code-based APIs.

Another investment priority for the firm is the intersection of AI and fintech, where Frenzel sees significant opportunity, as many financial institutions are still in the early stages of AI adoption.

This focus reflects how Kadan operates internally. At the firm, automation is woven into nearly every stage of the investment process, from deal sourcing and due diligence to portfolio management.

For instance, the company uses AI to track signals such as when a new founder has joined an accelerator or when another has left a notable tech firm to build a new startup in stealth. AI agents crawl an entrepreneur’s online footprint – from LinkedIn, GitHub, X, to Product Hunt launches. They also generate meeting memos and suggest questions investors can ask a prospective portfolio company.

After analyzing hundreds of data points, including a founder’s track record of exits and whether they started their firm solo or as part of a team, its AI model assigns each founder a score based on the likelihood of raising a next round or ultimately achieving a significant exit.

It also considers the amount of time an entrepreneur has spent in San Francisco, which Frenzel deems is “a big predictor of the network” that these founders have and their hiring capability in that ecosystem.

An AI-native investor for AI-native companies

Do all VCs need to be AI-native?