Money20/20: Hummingbird on the Path to AI Enablement

In an exclusive interview with FinTech Magazine at Money20/20, Matthew Van Buskirk, Co-founder and CEO of Hummingbird Regtech, discusses how the company is enabling data to be labelled at ease through the use of artificial intelligence. 

The Data Challenge

Hummingbird positions itself as an AI enablement company, addressing a critical challenge in the financial sector: the lack of well-labelled, clean training data. 

“From the day we launched, we saw that deploying effective machine learning or AI capabilities in this industry is incredibly difficult because we don't have clean training data. Instead, we have thousands of ad hoc spreadsheets -- a data scientist's nightmare,” says Matthew.

The company's solution involves extracting and transforming this disparate data into usable, clean formats, creating a feedback loop where decisions made within the platform organically label data, which customers can then use to train their in-house models. 

This year marks a significant milestone for Hummingbird as it begins deploying its own AI capabilities, leveraging developments in raw processing power from major technology providers.

Transforming Investigator Efficiency

“Historically, financial crime investigators spend about 90% of their time data gathering, cleaning, and doing paperwork, with only 10% dedicated to actual investigative work,” Matthew explains. 

“We're really trying to invert that ratio by having data models and AI capabilities prepare all the information for financial institutions.”

This is because, at their core, financial crime investigations are exercises in data sciences. Analysts aggregate data from dozens of sources to understand the activity that has been flagged and determine if it is nefarious. 

Traditionally this work has largely been done in spreadsheets but AI-based analytical capabilities are transforming the space. 

Matthew says: “Hummingbird reduces data gathering and cleaning time by around 70%, leaving much more time for the investigative value add of making intuitive connections and putting patterns together. This wouldn't be possible without accelerated computing, powered by the NVIDIA and Oracle platforms.” 

“We're trying to be the human-in-the-loop platform,” he continues. “The goal is to let humans do what they do best – figure out what's going on, apply their intuitive know-how, make decisions – and then let the computers run with it after that point.”

The platform's demonstration at Money20/20 showcased these capabilities, highlighting how seemingly simple automated validations and data processing can have profound implications for financial crime investigation efficiency. 

What might appear as basic text field operations and automated validations represents a significant advancement in investigation methodology.

Market Evolution

The company's growth trajectory has expanded beyond its original customer base. 

“Our customer base has historically been very large fintechs running in cloud environments, but we're now expanding to support global tier-one banks and other financial institutions,” Matthew notes. 

This expansion requires the company to operate across diverse technical environments to meet varying customer requirements.

Understanding the impact of these technological advancements requires deep domain expertise, yet the implications for the financial services industry are significant. 

He adds: “It's exciting to see this whole ecosystem of AI-enabled capabilities tackling these esoteric, behind-the-scenes, but critical aspects of financial services,” says Matthew. 

“The positive response to our demo shows that people understand what this technology means for the future of our industry.”

Explore the latest edition of FinTech Magazine and be part of the conversation at our global conference series, FinTech LIVE. 

Discover all our upcoming events and secure your tickets today.

FinTech Magazine is a BizClik brand