top of page

Juan Constain

Cofounder - COO/CPO

Quipu

MIT

MIT DUSP MCP 2018

Quipu

AI to democratize access to capital for microbusinesses

Location:

Colombia

Founded:

2019

Team Size:

4

Industry:

Artificial Intelligence (AI), Financial Services

Founders:

Juan Constain - COO/CPO
Mercedes Bidart - CEO
Viviana Siless - CTO
Eduardo Carrasquilla - CFO

About:

At Quipu, our mission is to unlock access to capital to unbanked and underbanked microbusinesses across Latam. We do this by developing disruptive AI based credit scores based on alternative and unstructured data. We've underwriten over 15,000 loans to 12,000 clients using our technology since 2022. Our lending customers are individuals who have a business and are unbanked or underbanked, are largely informal and 90% of them are rejected by traditional banks. Our scoring as a service customers are micro finance institutions, banks, fintechs, credit coops.

The Founder Files .png
OGMA Founders Image.jpg

Describe your startup. What is your startup's mission, and what problem are you solving and who are your customers?

At Quipu, our mission is to unlock access to capital to unbanked and underbanked microbusinesses across Latam. We do this by developing disruptive AI based credit scores based on alternative and unstructured data. We've underwriten over 15,000 loans to 12,000 clients using our technology since 2022. Our lending customers are individuals who have a business and are unbanked or underbanked, are largely informal and 90% of them are rejected by traditional banks. Our scoring as a service customers are micro finance institutions, banks, fintechs, credit coops.

The Founder Files _ Crimson Founders.png

Who are the founders, and what sparked the idea for this venture?

Mercedes and Juan met at MIT while pursuing their masters in city planning. Their background is mostly on economic development, public policy and microfinance. They spent years working on informal settlements across Latam before coming to MIT. Viviana was doing her postdoc at Harvard Medical School developing algorithms to detect injuries in brain scans and MRIs. Eduardo led innovation and finance projects at one of Latam's largest asset managers (Sura) for over 15 years.

Their knowledge about the financial inclusion problem from a different angle, gave them insights into how to solve it. The first idea Quipu pursued was a marketplace with its own community based and blockchain enabled currency. While implementing this idea as a pilot in 2019, they learnt that the biggest hurdle this businesses faced was access to capital.

What makes your product or service unique in the market?

We are the only technology company that has its own lending operation where we can test, iterate and launch disruptive credit scores with a real portfolio to then sell through our B2B business model. Our algorithms are based on years of studying informality and have deep expertise in machine learning and AI.

What milestones have you achieved so far if any, and where do you see your startup in the next 2-5 years?

We've underwriten over 15,000 loans to more than 12,000 clients in Colombia since 2022 using our technology and now are ready to scale allowing others to use our algorithms.

What defines your company culture, and what type of talent are you looking for?

We are untypical founders. We apply expertise from other fields in social sciences to the issue of financial inclusion. We pair this with disruptive machine learning engineers, data scientists and technologists that are passionate about advancing our mission. We are looking for young talented scientists that want to contribute to supporting the largest economy in Latam: unbanked and informal small businesses.

Are you currently open to investments or partnerships?

yes, we are extending our seed round, looking for an additional $1m USD.

How can interested investors, partners, or potential employees get involved or learn more?

What were the best and worst pieces of advice you received as a founder?

Best: prototyping means failing earlier so you can learn faster.
Worst: don't delegate in early stages.

bottom of page