How Aya got into investing
My entire career prior to VC was in highly-regulated spaces. During and after college I worked in the intelligence industry for the government, and when I transitioned to the private sector, it was to work for big banks. I was a coder, so a lot of what I was doing in those spaces was designing algorithms. But even within those highly-regulated industries, I was continually drawn to the innovation departments of the firms and agencies I worked at. And I was always engaged in entrepreneurial activities within them—starting newsletters, volunteering to lead committees, and creating innovative ways to enhance legacy algorithms, for example.
My last corporate role was at a bank in San Francisco. So I worked in a corporate environment by day; but after-hours, all my friends were in the entrepreneurial community. I learned a lot about venture from that particular community. The call was compelling, and I finally took the leap. I applied on Kapor’s website and—lucky for me—they reached back out. That was three years ago.
Key opinions
* Consuming foreign content can be an antidote to groupthink. I’ve noticed that while there are plenty of newsletters and podcasts for VCs, they can sometimes lead to groupthink. If I’m consuming exactly the same content as everyone else, it becomes harder to cultivate my own creative thoughts—let alone know if a thought I have is truly my own. To break out of that cycle, I often turn to foreign content. Since I read Japanese, I rely on magazines like Trendy and media like News Picks. I’m also exploring similar content from Chinese and African markets, possibly using AI tools for translation. Looking beyond the usual sources can provide a fresh perspective on the market.
* Onboard “Swiss Army knife” solutions. I'm a huge fan of tech tools that are everything-for-someone rather than something-for-everyone: The “kitchen-sink catchall” solutions tend not to do any one thing particularly well. One of my Swiss Army knives happens to be Harmonic; but whatever you choose, opt for technologies that are built specifically for the venture community and for the investor persona.
Key outlooks and predictions
* Foundations and impact funds will team up to provide more risk-tolerant, patient capital as traditional early-stage funds become more cautious. Right now, even the earliest-stage funds are playing it much safer: Some pre-seed funds are waiting until a startup is post-product, post-launch, or even post-revenue before they’ll commit. I’m noticing that foundations are stepping in to fill this gap, often providing patient, non-dilutive funding for capital-intensive startups, like those in climate tech or infrastructure.
Given this risk-averse climate, I also expect to see more collaboration between foundations and impact funds. Since they often share similar goals, this partnership could become more common. In fact, over the next decade, I wouldn’t be surprised if many VC firms evolve to become impact-focused—either partnering with foundations or having their own foundation arms.
* Near-future unicorns will solve for unexpected problems. In the imminent future, I think we’ll see venture capital shift toward founders who aren’t just refining existing solutions and workflows but are really digging into deep-rooted human challenges—problems people might not even realize they have, like chronic pain or fatigue. These will be the kinds of founders who focus on core human needs and come up with transformative solutions for issues that have been around forever but were once thought impossible to solve. I think that may be where the next wave of unicorns might come from.
My earliest investing challenges
I think my biggest early challenge when I joined Kapor occurred uniquely because I didn’t have a venture or venture-adjacent background: I hadn’t fully grasped how network-driven the industry is. In my life prior to VC, I was in front of a computer all the time, so I had no need to network or build professional relationships. When I started at Kapor, suddenly everything was about deal flow shares and ecosystem-building events. I’m a constitutionally shy person, so it took a good year-plus for me to feel comfortable getting out there and striking up conversations without knowing exactly where those connections might lead. It wasn’t easy to exercise that muscle, but I’ve since learned that I have a socially-confident quality to my personality, too.
The other challenge was gaining a clearer understanding of early-stage indicators. At first I was hyper-focused on hard metrics when diligencing companies or deciding whether to meet founders. But the early-stage game is fundamentally different from later- or growth-stage, where you have a lot of revenue-type metrics to base decisions on. Early-stage signals are necessarily nuanced. That’s one thing that drew me to Harmonic: It’s a data-driven platform, yes; but the signals it serves up support early-stage investors who just can’t rely on revenue signals. Does the founder have prior experience starting a company; and if so, what was the outcome? Do they have a strong technical background or prior experience in the industry they’re now building in? What’s the founder’s relationship with their co-founder, and how long have they known each other? There are so many ways to generate signals, and if I had to do it over again, I would prioritize more refined signal-gathering early on in my career.
One investment I’m excited about
Edtech is a space Kapor is increasingly exploring, especially in the realm of personalized education at scale. Without naming names, I recently sourced an edtech company I’m particularly enthusiastic about and took it to deal. The company is focused on test prep, and here’s what it’s doing differently: There are a lot of test prep platforms out there right now, the majority of which emphasize rote memorization through repetition or flashcards. But for test prep to evolve—cognitive psychology tells us this—we have to get beyond traditional memorization-based learning. We have to ask how students understand the material, and how they contextualize the subject matter.
Let’s say a test-taker answers a question correctly, for example. Did they answer it correctly by virtue of sheer luck? Because they comprehended part of the question and took a good guess on the part they didn’t understand? This company’s platform analyzes how people understand the questions they’re given, and leverages that intelligence to deliver the right types of resources to someone based on their learning style. They’re scaling a business model that way. That excites me because it will fundamentally change how people learn while democratizing test-taking for folks.
Using “controlled entropy” to source companies
Early-stage investors have to deploy a variety of strategies and channels to source startups. There are traditional co-investor networks, and I have regular syncs and deal-flow calls with co-investors in which we share the companies we’re evaluating. There are ecosystem events—happy hours for founders and investors, for example. Sometimes I attend these, other times I organize them. But I’m also thinking a lot about controlled entropy—a term used to describe the balance between order and disorder within a system. In my life as a startup investor, I think of controlled entropy as being in the right place at the right time—on purpose.
Here’s one example of this: I source a list of global unicorns that gets refreshed regularly. I conducted an analysis of U.S. unicorns to identify the funds that have invested in the highest number of unicorn companies in every sector. I put that data into a Harmonic list and noticed one particular angel who’d made an outsized number of investments in companies that became unicorns. I put that person through a people filter in Harmonic, did some research, and learned about an event he was hosting. So I made a business case to attend that event. Ultimately I wasn't able to meet him, but I met a lot of founders who wanted to meet him, which led to some high-quality deal flow. There are a lot of ways to leverage tools and signals to put yourself in environments of controlled entropy like this.
I don’t want to automate context-driven or intuitive decisions
I don’t use AI as much as some people might think: It can be a double-edged sword, especially for early-stage funds. There are a lot of AI tools out there that automate the repetitive tasks that don’t require much thought, and I’m in favor of that. But there’s too much nuance and intuition in how an early-stage investor operates to introduce AI into ultimate decision-making. I would never want to automate context-dependent decisions—for example, I wouldn’t automate a diligence process unless it was a first-level screening to determine whether a company is out of scope. When it comes to deep diligence, I’m only ever going to rely on human-first, human-centric decision-making.
That said, if I’ve got a bunch of inbound inquiries and I’ve already determined through an automated process to pass on 20 of those companies because they’re out-of-stage, I’m willing to send a bulk response created with AI automation. I might set up an if-then logic: If X, then send a pass email. That sort of decision-making I’m comfortable automating. But when it comes to early-stage investing, I hold the view that it’s better to err on the side of being overburdened with work than to miss a diamond in the rough because you’ve automated your workflow for efficiency.
