As a venture firm focused on artificial intelligence, Seattle-based Flying Fish looks for untapped areas where generative AI tools have not yet been built. But that’s been a challenge.
“The pace is crazy,” said Flying Fish principal Heather Gorham, speaking at a panel discussion about AI startups in Seattle this week. “White space gets filled really fast.”
New AI startups are popping up at a rapid pace, fueled by more than $14 billion in venture capital through the first half of this year. Meanwhile, large companies like Microsoft, Amazon and Google are in a race to integrate large language models into their core business strategies, making it difficult for smaller companies to contend for market share.
The panel discussion, at the TF Labs AI Summit, also featured Kevin Leneway, a software engineer at Pioneer Square Labs, and Larry Colagiovanni, a partner at Madrona Venture Labs. The conversation focused on tips for founders building AI startups and commentary on the AI startup market more broadly.
Here are takeaways from the discussion, moderated by Jonathan Blanco, founder and CEO of TF Labs. Comments were edited for brevity and clarity.
Founders: What is your data story?
Colagiovanni: “So much AI ends up getting commoditized over time that it’s hard to have actual AI be a differentiator, especially with generative AI. It’s easy for people to spin up applications, so we quickly glance over the AI slide. We ask founders: do you have access to unique or proprietary data? Are you bringing together data sources that are typically disparate, and that other competitors cannot? We look for foundational questions about your customer and business, then really try to unpack your data strategy. Is that something that will be a moat for you over time?”
Differentiate
Leneway: “One thing we look for is subject-matter expertise. I can go to ChatGPT and ask for a closing statement for some trial. But a lawyer knows day-to-day what you actually need. It comes down to fantastic people who are willing to break down walls and get this out into the world because they just love doing it every single day. You really do have to understand the nuances and workflow.”
Exciting crossovers
Gorham: “If you think about all this technology applied to finding new cures for cancers and diseases that we have no idea how to solve — that’s just mind-blowing. From a sustainability perspective, there are so many smart people applying AI to those problems, too. For me, that’s the most exciting space for me to spend my day-to-day, that crossover between science and machine learning.”
Beyond the demo
Leneway: “It’s so easy to put together a demo and landing page that says, ‘we’re lawyer-GPT, we can replace your law firm,’ or ‘we are sales-GPT, and we’ll do all your sales calls for you.’ Everybody can make a cool demo. If you’re going into this, you have to have a really clear go-to-market strategy and way that you can actually get some customer traction right off the bat.”
Competition is fierce
Colagiovanni: “If I was trying to convince someone not to build an AI company, the biggest reason is you are going to find a competitor — or five competitors. It is intimidating and scary. It takes a lot of tenacity, focus and drive to plow through that. If you’re not okay waking up every morning and reading about some new company that just raised a million dollars and adding it to your competitive list, I would strongly consider not doing an AI company. That’s just what it’s like right now.”
