From compliance to stablecoins to microbusinesses, fintech labs germinate next-gen uses for AI.
You wouldn’t think that the quipu—an abacuslike system of knotted cords used by the Incas for record keeping—would have much application to the breakneck adoption of artificial intelligence among financial institutions (FIs). But a Colombian financial-services company nurtured by Bancolombia Ventures harkens to that system. Quipu deploys AI-powered analyses of alternative data to determine the creditworthiness of microbusinesses.
Quipu’s work is just one example of the growing importance of AI to financial institutions. According to Statista, the financial sector “exhibit[s] one of the highest adoption rates across industries.” In fact, Statista estimates that in 2024, the financial services industry invested roughly $45 billion in AI technology. Concurrently, NVIDIA found that more than half of the companies represented in its global State of AI in Financial Services: 2025 Trends report view AI as “crucial to their future success.” Of the 600 financial services professionals surveyed, 98% of managers say that their organizations plan to increase AI infrastructure spending this year.
Many banks have already deployed AI to automate internal processes such as customer onboarding, credit scoring, fraud detection, and loan processing. Increasingly, FIs consider AI a pivotal tool for efficiency and cost-effectiveness in meeting evolving anti-money laundering and know-your-customer regulations.
As these innovations become more commonplace, some banks may wonder what’s next for AI? That’s where innovations arising from the world’s best fintech labs come in.
AI capabilities continue to mature. Enhanced AI capabilities will help FIs generate new business value, but only if those institutions follow the advance from AI to generative AI (Gen AI).
The term “artificial intelligence” is used for technologies that can perform tasks previously requiring human brain power. Relying on historical data and rules-based systems, these capabilities recognize patterns, understand language, and detect anomalies—notably, the types of anomalies that can indicate fraud.
Gen AI is a specialized branch of AI that exceeds content analysis to actually produce content. Gen AI can write. It can simulate human conversation. It can code. It can generate images and videos.
The difference between AI and Gen AI can be seen in chatbots. Imagine this: A customer asks a chatbot, “Why was my credit card application denied?” An AI-powered chatbot may return a list of common reasons for the bank to deny credit, followed by a customer-service phone number for the user to call. A Gen AI-powered chatbot may respond with, “Your credit card application was denied because your credit score is too low. Your credit score is too low because a $2,000 write-off appears on your credit report. This write-off seems to be related to an auto loan from ABC Motors. Repaying this debt will help you improve your credit score. You may want to contact ABC Motors to settle this debt. Consider negotiating a ‘pay-for-delete’ arrangement.”
“Loan sharks were these businesses’ only solution. We’re an alternative to that.”
Mercedes Bidart, CEO and Founder, Quipu
Chatbot improvement is just one way Gen AI can improve business for FIs. It can study customer data to more closely tailor marketing strategies and financial services to individual needs. It can improve loan and investment strategies by generating “what if ” scenarios to help banks chart, for example, how changing interest rates affect customers’ willingness to take out new loans, and customers’ ability to repay those loans. And, as the innovations discussed below indicate, AI and Gen AI can help banks and their clients hasten international trade. They can help spot and stop previously unknown threats to bank infrastructures and data. And they can provide a financial lifeline to the underbanked.
Bancolombia Ventures partners with startups, focusing on topics such as fintech, climate-related technology, and cybersecurity. One of those startups, Quipu, has developed a new credit-scoring system tailored to what Bancolombia calls the “informal” nature of business in Latin America.
According to El Pais, a leading Colombian newspaper, close to 95% of all businesses in that country are microenterprises—defined as operations with 10 employees or less. While employing 65% of the Colombian workforce, these organizations tend to suffer from “business dwarfism,” or an inability to grow. Why? They lack access to capital. Traditional credit scoring methods paint them as a bad risk.
This is a problem that Mercedes Bidart, Quipu CEO, is trying to solve. The MIT graduate notes that most microentrepreneurs in the country operate as freelancers. “They have their digital wallet or bank account as a person, not as a business,” Bidart says. “They come in for an SME [small or midsize enterprise] loan at the bank, but they won’t get that. There’s no information about their business behavior.”
The Quipu system finds new ways to detect business value. It looks at business location, social media posts (including videos, pictures, and customer comments) and other nontraditional sources of information to determine business health. Even Google Maps can indicate whether a business is growing—showing, perhaps, the physical expansion of a home-based garage over time.
Quipu uses this information to develop its own credit scores for microbusinesses. Potential new clients are often referred by Bancolombia, from its pool of declined applicants. Quipu has offered many of these microbusinesses loans ranging from $100 to $2,000—for a total of $3.5 million in loans granted over the past 18 months. While these are personal loans, rather than business loans, Bidart believes that these small infusions of cash will help some businesses grow to the point where they eventually qualify for more-traditional SME loans.
“The people we serve—before us the only financial solution they had was the predatory lender. We have loan sharks. They charge abusive interest rates, and they’re violent,” Bidart says. “They operate from Mexico to Argentina. In Colombia, loan sharks were these businesses’ only solution. We’re an alternative to that.”
Let’s take look at innovations arising at other fintech labs around the world.
