Can AI help female founders get funded?

Last fall I had the opportunity to pitch a well-connected Silicon Valley angel group and the great fortune of making it through due diligence with them. Then, after many hours spent working on a potential investment, I learned that they would not be investing—even though my company was rated as having potential for “high return and medium risk.” In conversation with the lead investor we were working with, she commented that one of the reasons she was passing is that “it is harder for women to raise a Series A.” Meaning that she’s invested in women founders before and that her investment struggled to get a return—she seemingly calculated the founder’s gender into her investment decision-making process. This experience got me thinking: What if AI had been used to support this investor’s decision-making process? Just like AI is built on pattern recognition to make predictive models, this investor had noticed a pattern (women have more difficulty than men raising a Series A) and used that pattern recognition to guide her decision-making. Would AI level the playing field by eliminating the discrimination I faced last fall or simply reinforce those structures? What the AI experts say I recently sat down with a number of AI experts to better understand how AI could help or hurt founders like me, who have been historically underfunded in the largely relationship-based business of venture capital (VC). When asked this question, reactions were mixed with one expert even telling me “I fundamentally disagree that AI should be used” when making investment decisions. AI expert, advisor, and founder of Almma Health, Ashley Kienzle pointed out one of the problems with utilizing AI to analyze data sets for insight into future predictions is that “biased data is just going to create biases.” As Ashley explains “there’s even bias in female-run VCs.” She suggests even if gender is removed in steps like pitch deck screening to build AI models, these data sets can be fraught with bias. Simply looking at other pieces of data, like occupations, can accidentally enforce bias when 65% of STEM positions are held by men. In her opinion it comes down to “what are you trying to measure and predict?” If you’re trying to predict the next “unicorn,” there isn’t enough unbiased data to build a predictive model. Martin Gedalin, founder of Rings XRM, offered a more analytical perspective. Martin’s team is working on solutions for investors with complex use cases for larger firms and family offices. Their process is built on key business metrics and a mix of public and private data. They do not include the founder as a key data point. This, in theory, should level the playing field, that is, once a business reaches certain financial metrics, gender and identity of the founder will be irrelevant to their model. Jennifer Kloke, VP of artificial intelligence for Autonomize, had a measured optimistic perspective. “It is very possible that a well-built model that thoughtfully avoids basic bias might actually open up more opportunities for female founders where the current state is “gut feeling” for the VCs.” Jennifer goes on to point out that “AI models will not replace the need to talk to people” and likely the next wave of AI tools for investors will be selectively built around analyzing financial models or market data to guide in decision making.Lauren Washington, the founder of Fundr, has an AI based tool to help investors optimize their process and make better investment decisions by eliminating bias in their decision-making. In order to eliminate bias, they specifically take into account the founder’s gender, race, and other demographics. “What I am seeing is a shift in the industry right now as people are using that data to highlight disparities and pinpoint places to change it.” AI as a tool Each AI expert I spoke with emphasized the need for a large volume of high-quality data to build generative AI models that will predict future success, which begs a bigger question: Is there even enough data to predict the likelihood of success for any one investment? As it stands today, this seems unlikely. However, the likelihood that AI may be utilized to analyze financial data or market potential in due diligence is very high, and already happening. On its surface, AI is a tool. And like any good tool it all depends on how the user utilizes it. If software is built without honest recognition of the biased history in the VC industry, it is probable that pattern recognition-based models may inadvertently pick up on the patterns of bias in past deals. While AI may not immediately change the way founders get funded, it is possible that unbiased analysis of markets and financial projections could highlight overlooked opportunities for investors. If this will level the playing field for underrepresented founders like me, I am all for it. Maureen Brown is cofounder and CEO of Mosie Baby.