00:00Our next guest's investing thesis centers on what she calls the, quote, AI-pilled organization.
00:05She writes and speaks on what this shift means for the structure of startups,
00:09where new opportunities lie in the next generation of category-defining companies.
00:13Floodgate partner and co-founder, Anne Murakow, joins us now.
00:18What does that mean, AI-pilled?
00:22It's a good question.
00:23What we're seeing these days is that there's a new way in which companies are going to be built
00:30and organizations will be run.
00:34And when we say that organizations are AI-pilled,
00:38most of the time people are thinking about engineering organizations using tens, hundreds, thousands of agents at a time.
00:46But we're starting to see actually what I would call almost software factories being built
00:52so companies are not just building their own products using software,
00:59but they are actually building the factory that will actually build the software using AI.
01:06And that means everyone from sales, marketing, strategy, they are individuals also using AI.
01:13And we think that's really fascinating today.
01:16There's two different things that jump out at me.
01:18One is speed and one is size.
01:20Like in the news flow, how many companies have been formed in the last year alone
01:25by alumni of any frontier lab or any software company where it's just a few individuals
01:29and very quickly they reach critical mass.
01:32Is that kind of what you're talking about?
01:34Some of it is about the speed, but I would say what's really interesting today
01:39is that the learning cycles for these companies is actually becoming extraordinarily fast.
01:47And so when we've talked to over 30 companies that are from three-person startups to public companies,
01:57they include AI companies like Vercel, Decagon, Harvey, and Ramp.
02:03The big difference is that they're learning so much faster because every part of seeing the data
02:11and even acting upon that data and then learning from what was actually done
02:18is happening so much faster because it can be done agentically.
02:23And in Fluggate's sort of pursuit of trying to understand what's happening in the real world,
02:27you went out and surveyed or interviewed quite a large number of individuals
02:31and quite a large number of companies, big and small.
02:34Talk us through why you did that, the methodology, and net-net your conclusion.
02:40Yeah, and this work is not finished by any means because what we're finding is that
02:46every month everything starts to change.
02:49I was actually really inspired by one of our portfolio companies where the CEO,
02:54maybe like a year before everyone really woke up to the fact that coding was changing,
03:00came to me and said, I'm having an existential crisis of how our companies will be run.
03:05And when he said that, I said, well, I'm going to have to figure that out.
03:09And so I went and talked to the most AI-native companies and found that not all AI product
03:17businesses are actually using AI internally.
03:20So then I decided, well, I need to know who's actually using AI internally outside of engineering.
03:26And some of the companies that really jump out to me are companies like Ramp,
03:31where they are really enabling all of their employees to leverage AI in new ways,
03:39from a salesperson to finance across the board.
03:45Vercel is another company that we saw that was very far ahead in terms of how they were also leveraging
03:51AI
03:52and enabling all of their employees to generate AI agents and use them.
03:59So, and, you know, fast forward to present day and what happens next.
04:03How does all of that work, which is ongoing, determine where you guys invest or don't invest at Floodgate?
04:11Yeah, we believe now that, first of all, we've used this research to actually generate our own agents
04:19internally influence the way we work.
04:21But the second piece is that there's a lot of holes, right?
04:24So agents really need, in order to operate, and for employees to actually really leverage AI,
04:31there's a certain amount of context and knowledge that you have to digitally document.
04:36That is still not done extraordinarily well in most organizations.
04:41And then how do you make a regular employee comfortable using AI is still a big question.
04:48And companies like Ramp can actually build harnesses and software that enable their employees to do so,
04:55but a three- to ten-person startup doesn't have that capability.
05:00And so there's still a lot of infrastructure and software to be built,
05:04and that really creates sort of opportunities for companies like Floodgate to come in
05:10and invest in founders who see that future as well.
05:13And it sounds like at the earliest stages, right, you know,
05:16Ramp is a pretty mature company at this point, relatively speaking.
05:21Yes.
05:22And this is why Ramp has the resources to be able to build those harnesses internally.
05:28And we're seeing some of these companies that are fairly large,
05:31actually hire AI internal tooling teams, but that's not going to last forever.
05:39Or not every company is going to be able to hire those individuals.
05:41And so we believe that there will be infrastructure that is built.
Comments