ELSA's founder and CEO, Vu Van, was recently featured by EXAME, one of Brazil's largest business publications. The interview covers her path from struggling to speak English at Stanford to building a company now used in 120 countries. It also covers ELSA's expansion into Brazil and the growing role of its corporate business — the part most HR and L&D leaders will recognize from their own teams.

Below is a summary, with a link to the original piece at the end.

A Language Gap That Only Appeared Under Pressure

Van grew up in Vietnam and built a strong international career. She worked as chief of staff to the CEO of Maersk in Denmark, then completed an MBA and a master's degree in education at Stanford. By most measures, her English was already good. She had studied it for years and scored well on tests.

That preparation did not carry over into live conversation. Once she had to speak up in a room of native English speakers, she found it difficult to be understood and began to lose confidence. The issue was not vocabulary or grammar. It was the gap between studying a language and speaking it under real pressure. She noticed the same pattern in other international students around her: years of formal study, but hesitation the moment they needed to speak.

That gap became the foundation for ELSA. Van founded the company in 2015. ELSA stands for English Language Speech Assistant. The starting question was simple: one-on-one conversation practice with a human coach cannot scale to billions of learners, so what can AI provide instead?

Ten years later, ELSA has more than 100 million users across 120 markets, has raised approximately $60 million, and has worked with more than 1,300 organizations.

Building an AI Company Before AI Was Widely Trusted

In 2015, describing a company as "AI-powered" did not generate the same investor interest it does today. Van says the team had to demonstrate that the technology actually worked, rather than rely on the term alone. Her co-founder, speech-recognition researcher Xavier Anguera, contributed the technical foundation. Van contributed the business and education perspective. Together they built an early version that could listen to a learner speak and give feedback in real time. That was enough to prove the concept could become a real product.

The company's financial approach has also changed over time. Van describes the current phase as building an efficient and sustainable business, rather than pursuing growth at any cost. Investment is now focused on three areas: AI and product development, the corporate business, and international expansion, with Brazil as a key part of that expansion.

Why the Shift Toward Corporate Clients Matters

ELSA began as a consumer app with a freemium model, and that model still exists today. But according to Van, the corporate side of the business is now growing faster than the rest and is receiving more of the company's attention.

Her reasoning is directly relevant to HR and L&D leaders at global companies. English proficiency affects sales performance, customer service, leadership visibility, cross-border collaboration, and overall productivity. An employee who writes clear, accurate emails but struggles to speak confidently on a live client call does not have a vocabulary problem. It is the same gap Van experienced at Stanford, appearing in a workplace setting instead of a classroom.

ELSA cites clients such as TELUS Digital, which employs approximately 78,000 people across 31 countries, and NashTech, where the company says its English program is connected to professional development and, in some cases, promotion requirements. Van describes ELSA's goal as building an AI-based communication platform for global companies, rather than simply an app for studying English.

Why Brazil Was Chosen as a Priority Market

The interview frames Brazil as more than one market among many. It is where ELSA has established its regional hub for the Americas, based in São Paulo.

The reasoning is straightforward. Brazil has a population of more than 200 million people and a large workforce. A study by Pearson and Opinion Box, based on responses from more than 7,000 Brazilians between February and April 2025, found a meaningful gap in English proficiency. Among Brazilians who report some knowledge of English, only 13% describe themselves as fluent, 38% describe themselves as intermediate, and 49% describe themselves as basic. Separately, 29% of respondents said they have no contact with any language other than Portuguese.

ELSA already had a large user base in Brazil before this expansion. The regional office is intended to build on that consumer presence and grow the corporate side of the business. The team is staffed to work across Portuguese, Spanish, and English, and is also expected to support other countries in Latin America. Danielle Alfieri, ELSA's Executive Director for the Americas, is quoted in the piece: "We don't see Brazil as a short-term initiative; we're building the foundation for a long-term regional presence."

ELSA has not disclosed how many Brazilian companies it currently serves, and has not set a public revenue target for the country. The company states only that it already reaches millions of Brazilian learners and continues to add corporate clients.

Why Not Just Use a General AI Tool Like ChatGPT?

This part of the interview is especially relevant for HR leaders evaluating language-learning tools. General AI assistants can now hold a conversation in English, so the natural question is why a purpose-built platform is still needed.

Van's answer centers on data, not novelty. Many companies can now build a basic AI chatbot. What ELSA has built over the past decade is a large set of speech data from learners across different countries, accents, proficiency levels, and goals. This data allows the platform to identify specific error patterns and personalize feedback, rather than offer generic corrections. Van refers to this approach as hyperpersonalization: using a learner's own history with the platform to shape what they practice next.

Van also acknowledges that this advantage is not permanent. She notes that the product that made a company successful five years ago is not guaranteed to be the product that keeps it successful five years from now, and that generative AI has advanced faster than most people in the industry expected.

The Core Takeaway

The most useful part of this interview is not the funding numbers or the user counts. It is that Van's original insight — that studying English and speaking it confidently under pressure are two different skills — is the same challenge HR and L&D teams manage every day across global organizations, repeated across every market, meeting, and manager where English is not a first language.