Wenyi Cai and the question that took her to Colombia

Wenyi Cai had already helped run a Silicon Valley startup that eBay wanted to buy. The next move might reasonably have been another startup, another California office, another conversation about what the internet would do next. Instead, she went to Colombia to ask people about their lives. Before Polymath Ventures could build a company, it needed to understand the people a company might serve.

That choice has become the organizing idea of her career. Cai builds and invests in businesses for Latin America’s emerging middle class, a population whose everyday needs can disappear beneath the excitement of a technology pitch. Her interests include how people earn, shop, borrow and keep a small business working. The questions sound ordinary. Getting useful answers has occupied her since 2012.

Her route there involved applied physics, quantitative trading, consulting and consumer software. It also involved leaving graduate school and putting her own money into an experiment in a country where she would have to learn the language. There is a pleasing inconvenience in this résumé: every time it appears to settle into a recognizable profession, she changes the subject.

A shopping engine, and a different destination

Cai was born in China and grew up between China and the United States. At Harvard College, she studied philosophy and physics and graduated cum laude. Before her business career, she published applied physics research and helped build a quantitative trading hedge fund. The combination gave her experience with both abstract problems and the uncomfortable business of finding out whether a model works.

She later worked for McKinsey across the Middle East, Africa and South Asia. Based in Dubai, she saw mobile networks create new kinds of businesses and services. Phones and SIM cards were concrete evidence of technology changing commercial life. The interesting part was how those changes reached people, beyond the technology itself.

In Silicon Valley, she became chief operating officer of Milo.com. She led business development with companies including Microsoft and eBay, while overseeing product and engineering. Milo helped shoppers find accurate inventory and prices at nearby stores. It was a digital company whose usefulness depended on a very physical fact: whether the thing you wanted was actually on a shelf.

When eBay announced the acquisition on December 2, 2010, Milo’s retail relationships covered approximately 50,000 stores across all 50 American states. The idea was to connect online research with shopping in the world outside the browser. For Cai, the work supplied experience in turning software into a service people could use, and in coordinating the people who made it happen.

Then came the departure. While studying at Harvard’s Kennedy School, she began recruiting a team for a project focused on Latin America’s emerging middle class. She left the program to pursue it. The practical opening was a round of interviews with hundreds of people. A business idea would have to survive those conversations before it deserved a company.

A summer project acquires a future

Polymath dates its first ideation project to June 2012 in Medellín. Its own history describes ten recruits working for three months, producing the concepts for Táximo and Aflore. The beginning was small enough to fit into a summer. Its consequences would be considerably less considerate of the calendar.

Cai supplied the initial seed capital. That decision let the group concentrate on its first experiments while she was still a newcomer to the market. Building a venture studio also meant building the means to keep building: a team, a method and access to money. Each new company would need more than an attractive idea.

Her co-founder and finance counterpart, Carlos Fernández de la Pradilla, brought private equity and entrepreneurial experience. The studio’s work developed in Bogotá. In 2014, Polymath launched Aflore, Autolab and VincuVentas. It entered Mexico in 2017 and launched Elenas in 2018. The geography widened alongside the collection of problems the team was trying to solve.

Those dates describe institutional progress. Cai’s account of the work is more revealing. In a reflection published nine years after Polymath began, she acknowledged that the firm had started before several conditions were ready. Widespread smartphone adoption, early capital and experienced startup talent could not simply be wished into place. An early start bought time to learn. It also bought a long stretch of difficulty.

She described the limits of working like a scientist alone in a laboratory. Other companies eventually supplied logistics and financial infrastructure that Polymath’s ventures could use. Other venture studios provided peers. More investors gave promising companies access to capital. A builder might enjoy starting from first principles; rebuilding the entire commercial world is a rather demanding side project.

The customer gets a vote

The studio’s method gives that initial listening a place in a longer process. Ideation looks for unexpected insights and tests concepts. Validation builds a minimum viable product and examines whether people want it, whether it can work and whether they will pay. Product-market fit requires further building and testing, sometimes including major changes in direction.

Later stages address the economics of growth and expansion. The sequence matters because a product can be appealing without being a workable business. Polymath’s current studio site says it has prototyped more than 150 business concepts. The number describes experiments, rather than 150 operating companies. Much of the work takes place before anything merits a launch announcement.

Cai’s examples stay close to existing behavior. Elenas works with social commerce. Autolab applies technology to auto repair. The businesses have different customers and different operations, yet each starts with activities that already occupy people’s time and money. The studio seeks an improvement that makes sense within those routines.

Her thinking about China follows the same logic. In a 2023 analysis of digital payments, she and her co-authors examined how online services grow out of distribution, habits and infrastructure. They emphasized the work of bringing offline merchants online, and the differences between a centralized country and a region of separate markets. A precedent can be useful without becoming an instruction manual.

For Cai, local knowledge also belongs in investment decisions. In a 2022 interview, she argued that foreign investors might misunderstand businesses such as Elenas, whose model responds to Latin America’s established direct-sales culture. She also questioned the pressure to expand into Mexico prematurely. A larger addressable market looks splendid on a slide. A founder still has to operate there.

The useful part of being wrong

Cai talks about failure as something an ecosystem needs to learn from. Her experience in Silicon Valley exposed her to a culture that collected lessons from unsuccessful attempts. She wants those lessons to remain available to the people building the next companies, rather than disappear with the previous venture.

“What’s really insightful about failure is that you learn a lot through it.”

Wenyi Cai, 2022

The useful word is “through.” A failure requires examination. Otherwise, the next attempt inherits the same assumptions in a fresh presentation. Cai’s argument gives experienced founders a place in the ecosystem even when their earlier companies did not produce the desired result. Their experience can still change the quality of the next decision.

She applies that willingness to revise to Polymath itself. Her 2021 reflection argued that a studio needs allies and peers. It also described employees who left and later returned, and analysts who became venture co-founders. Careers inside this kind of organization need room to change. The company’s knowledge lives partly in relationships that extend beyond its current staff list.

Her view of talented people is unusually expansive for a business discussion. Challenge, mentorship, meaning and friendship all appear in her account of what people seek from work. That creates a different question for a CEO: how to make an organization where people can develop, stay connected and contribute across changing roles. Recruitment supplies a beginning. Relationships have a longer job to do.

After the pitch, the arithmetic

In 2023, Cai described Polymath responding to a difficult funding environment by pausing new company creation and supporting its existing portfolio. She stressed spending discipline, the economics of individual customers and the risk of relying on paid acquisition. If a company’s growth slows the moment it stops buying attention, its earlier growth deserves a closer look.

The firm also announced a dedicated seed fund that year, extending its work to external investments as well as studio builds. Pacto, a payments and point-of-sale business for Mexican restaurants, became its first direct investment. Polymath’s current fund focus includes fintech and software designed for particular industries. Its website also presents a private credit business offering asset-backed financing.

That broader set of tools remains connected to the question Cai took to Colombia. A small business may need better software, a workable payment service or access to financing. The appropriate response depends on its actual problem. Cai’s work has developed into an organization able to approach that question from several directions.

In February 2025, she was listed on the FII Priority Miami program for a discussion about new financing models and innovation in emerging markets. The program identified her as Polymath’s founder and CEO and a venture partner at BAI Capital. Her path continues to connect company building with the capital that makes it possible.

The enduring ambition is to create more economic opportunity for Latin America’s emerging middle class. It is a large goal pursued through specific commercial decisions: whom to serve, what to test, when to hire and when to expand. Cai began by asking people about their lives. Years later, that remains a useful way to decide what deserves to be built.