As the managing director of the AI2 Incubator, where she plays an active role in investing in the next wave of AI-first companies, Yifan Zhang shares a nuanced view on the pace and impact of AI innovation in the workforce.
“There’s a lot of hype around AI, and I do think that it is going to take longer for AI to really deeply change most areas of work,” she says. “It’s going to start slow and then come all at once, but at the end of all that, I hope that our work becomes more strategic versus more execution oriented.”
That’s one of her many insights in this episode of Shift AI, a show that explores what it takes to adapt to the changing workplace in the digital age of remote work and AI.
We discuss Yifan’s early successes as a startup founder and explore her new role as a leader in the Seattle technology ecosystem incubating young AI companies and helping them grow.
Listen below, and continue reading for highlights from her comments, edited for context and clarity. Subscribe to Shift AI and hear more episodes at ShiftAIPodcast.com.
Background and early experiences: I am an immigrant. I came from China to the U.S. when I was seven. Grew up in Indiana, of all places — not a lot of other Chinese immigrants — and then went to Harvard. When I was in school, I definitely did not know anything about entrepreneurship. I didn’t know it was possible to go out and just start a company. But I fell in with a group that ended up all becoming either entrepreneurs or VCs. It was the Harvard Entrepreneurship Forum which was the first iteration of many entrepreneurial groups now at Harvard. So I started my first company, GymPact, basically right after graduation. I was 22 when I started that company.
Early startup experience: GymPact was based on a paper I wrote my senior year at Harvard around financial incentives, specifically negative incentives and the ability to motivate really hard behavior change. So it was a fitness app, and became one of the top fitness apps back in the day, back in the 2012, 2015 era. We were the app that charged you for not exercising and paid you if you did. So we use financial motivation to track your exercise and verify to really create an incentive system to drive behavior change. Now, looking back, people think it’s a really obvious use case, but I think at the time it was seen as really crazy and out there. Mobile phones were new, mobile apps were new, and the idea of using financial incentives like that attracted a lot of attention.
A focus on housing: I started my second company, Loftium, which is an affordable housing startup. We help people become homeowners by creating part of their home, an income-generating short-term rental so they could save up a down payment and become a homeowner within three years. We call it ‘host to own.’ That company I started here in Seattle after traveling for a bit and living in different cities and deciding Seattle was a really great base to start another company. Loftium raised almost $30 million in venture capital and then put together almost $100 million in asset-backed debt to be able to actually own real estate across the country. And we sold that company to Flyhomes here in Seattle early last year, in 2023.
AI2 Incubator’s unique approach to startups: Right now, there’s a lot of “out-of-the-box” solutions in AI, especially after the launch of ChatGPT, but we think that the out-of-the-box solutions are really only good for 80% of what is acceptable to an actual customer. It’s really easy to build a dazzling prototype right now. It is very hard to take that prototype to something that’s actually usable, like an MVP. AI2 Incubator is really looking for those use cases where we think we can help a company go from 80% up to 90%, 95%, whatever acceptability looks like in a startup’s timeframe. So that’s less than two years on a pre-seed or seed budget.
The three T’s of company building: I have talked at the incubator with our companies about what I call the three T’s. What investors look for is team, traction, and technology, and depending on your stage, whether you’re pre-seed or seed, you need to check off one, two or three of these T’s. It also kind of depends on how hyped the VC cycle is. I think right now, given that VC is much tougher, including in AI, we’re encouraging our founders to be able to check off at least two of these T’s and ideally make progress on the third T, but you can choose which T’s you’re checking off.
Choosing the AI2 Incubator over standard investor roles: I actually hadn’t been thinking about going over to the investor side of the table, but I found the incubator model very interesting. There is a lot of room for me as a founder to come in and help the companies and work very directly with the companies versus being a standard investor, where you just chase a deal, select the founders, and then really you’re not getting to spend that much time on the building process. I really like the incubator process and the fact that AI2 Incubator is not a cohort-based program. We think of ourselves as a co-founder that comes with a $500,000 check. We put in both money, our own time and expertise, and I personally was very attracted to that.
The future of work: I think the future of work post-AI is going to be strategic. There’s a lot of hype around AI, and I do think that it is going to take longer for AI to really deeply change most areas of work. It’s going to start slow and then come all at once, but at the end of all that, I hope that our work becomes more strategic versus more execution oriented. There is so much of our time that’s spent on things that take longer than they should, and we try to automate and optimize as much of it as possible…and hopefully there’s still a role for us. I could see that, and I would be very happy with that type of world.
Listen to the full episode of Shift AI with AI2 Incubator Managing Director Yifan Zhang here.
