‘Nightmare’ Pitch Meetings: One Female Founder’s Long Road to Funding Her AI Startup

Growing up in Alaska, Heather Shoemaker always felt like the ability to speak a foreign language was a super power. That drove her to study French and Spanish and earn a bachelor’s degree in linguistics. While working as an interpreter and bartender after college, she was driving around Seattle when she heard a report on NPR describing Java programming as the future of technology. She decided on the spot to go back to school and learn a different kind of language: coding.

Shoemaker picked up an engineering degree and spent the next decade as a software developer reformatting source code for companies that were expanding operations internationally and needed their software to support multiple languages. She realized that the the biggest hurdle was not the software. Companies with a global footprint and a mostly English-speaking workforce needed more help providing customer support. With artificial intelligence and her skills as a developer, Shoemaker saw an opportunity to grant her super power to everyone.

In 2017, she launched Language I/O, an A.I.-powered SaaS platform that provides real-time, multilingual customer support in more than 100 languages.

Shoemaker was used to being the only women in the room, from her graduate school classes and working in A.I., where women make up only a quarter of the industry. As a female founder and CEO, she felt like she was in uncharted territory when in 2019 she went searching for investors for her bootstrapped company, which is based in Cheyenne, Wyoming.

Less than 2 percent of enterprise software companies are founded by women, and despite generating higher returns, startups led by all-female teams struggle when it comes to securing venture capital backing. During the first half of this year, women received only 2 percent of funding and accounted for only 6.9 percent of deals, according to Pitchbook data. Shoemaker assumes the number must be even lower for female founders who are software developers. Still, knowing the statistics did not prepare her for the reception she received at dozens of unsuccessful pitch meetings. 

“I talked to so many people that were just a nightmare,” she says. “I don’t know why they think what I’m doing is easy.”

Shoemaker could tell that the VC firms were not used to seeing a woman, especially one with technical skills as opposed to an M.B.A. During one meeting in Silicon Valley, a partner learned that Shoemaker was pitching hard tech and had written the original code herself. They asked, “How hard can it be to replicate what you coded?” Shoemaker says–insinuating that if a woman programmed the software, it must be simple for any other developer to copy, making it less valuable.

Another male investor, who passed, told her, “If you were an older guy with a long beard, an A.I. pitch would be a lot easier for me. That would just be more normal. We’re not used to this.“

Out one morning on her daily run, Shoemaker dreamed up a plan that reached farce-level proportions. Life would be so much easier if Language I/O had a male founder, so why not pretend to be one? She considered shortening her name to Heath, donning a beard on Zoom meetings, and went so far as to research voice-modulation software. Just before her fundraising efforts transformed into a Shakespearean comedy, Shoemaker decided against it.

“I thought about it a little bit longer,” she says, laughing. “You don’t want to start a relationship with a VC lying about who you are.” 

That dismissive attitude towards women working in A.I. is a problem not just for the female founders getting boxed out, but for society at large. By 2025, A.I. will have created 97 million new jobs, according to the World Economic Forum. As algorithms become further integrated into everyday life, the people writing the code will need to be reflective of the population.

The alternative–a programming monolith–produces biased results, even unintentionally. Gartner estimates as many as 85 percent of A.I. projects deliver erroneous outcomes due to bias in data, algorithms, or the teams responsible for managing them. That means companies looking to leverage A.I. will end up with faulty products.

For example, Shoemaker explains, if A.I. translates the word doctor from a gender-neutral language to a gendered language, the word that the program will almost always choose to give the user will be the male form while nurse will be translated to the female form.

“It drives me insane,” she says. “And now, because it’s not always gender binary, we need to start thinking about gender-neutral terms; nobody’s even broached that topic yet.”

After dozens of rejections, Shoemaker had considered giving up on venture capital, but in October 2020 she found a bit of success closer to home raising $500,000 in Wyoming. But she needed more capital. A Jackson, Wyoming-based advisor connected Language I/O with the East Coast angel network Golden Seeds, which focuses on women-led startups, and the Boston-based investor Eric Schnadig, who told Shoemaker he was going to introduce her to a so-called “super angel.” Shoemaker did not know what that meant, but was willing to meet with any potential backers at this point. In Boston, she finally met someone who really heard her pitch.

Bob Davoli, the founder and managing director at Gutbrain Ventures, reacted much differently than the other investors Shoemaker had pitched. “He immediately got the market opportunity and wasn’t arguing with me about market size,” she says. “He was not under the false impression that Google is going to do exactly what we do.”

Davoli, whom Shoemaker described as a feminist, was not your typical VC. Known for a hands-on approach and a hobby moonlighting as a singer-songwriter, he had a 30-year track record as an investor, and had landed on the cover of Businessweek in 2000. His portfolio comprised primarily enterprise software platforms, including startups focused on A.I. Shoemaker says finding an investor with that depth of sector knowledge was crucial and made the pitch easy.