The early testers hung up. That is how Jennifer Zhang's first talkbots met the public: a person picks up the phone, hears a machine, and ends the call.
The idea behind them had started somewhere much smaller. Zhang has said she watched her mother struggle with online customer service, stuck in a system built for fluent English speakers. In Southeast Asia, over 630 million people speak more than 1,000 languages and dialects, and she thought a machine should adapt to them, not the other way around.
Beyond the hang-ups, Southeast Asian speech mixes languages in the same sentence, and Singlish is its own challenge.
Her route to that idea was unusual. She majored in film at Keio University in Japan and Zhejiang University in China, then took a Master of Public Policy at the University of Southern California. In the US she was a columnist for an entrepreneur magazine, and her interviews included Mark Zuckerberg and Jessica Alba. She then spent nearly three years in venture capital, as Managing Partner of GCC Capital in Beijing and General Partner of Cybernaut Zfounder Ventures.
What almost nobody sees is why she left. She felt venture capital was too passive. She wanted to build instead of watching others do it. In early 2019, with two partners, Jianfeng Lu and Tony Zhu, both researchers in language technology, she co-founded Wiz.ai and chose Singapore as the base.
The first stretch was trial and error. Beyond the hang-ups, Southeast Asian speech mixes languages in the same sentence, and Singlish is its own challenge. The company worked at making the bots sound and listen like locals. In 2020 it raised US$6 million in a pre-Series A led by GGV Capital, and saw 200 percent growth that year. Hillhouse Capital led a US$20 million Series A, and the team grew from three people to almost 300 across 17 countries.
In November 2025 the company closed a Series B in the tens of millions of US dollars, led by SMBC Asia Rising Fund.
The cost of the move from investing to operating is not documented, and we will not guess at it. The idea she carried from her mother, though, survived every redesign. A founder can borrow a problem from a family kitchen and still build a company around it. The trick is to keep listening to the person who first had the problem, even after the bot has learned to talk.