Timing, Conviction and Insights on Startups and Fundraising (feat. Krish Ramineni)

All right. Okay. Welcome to the show, Krish.

Hey, KP. I'm excited to be on the show.

So, Krish, you're joining us from Seattle, right? And you said earlier before we got started that you kind of shuttle between the Bay Area, Seattle, and back and forth. How are you liking Seattle lately?

I'm enjoying the weather. It's much nicer now in the summer. I think the sun sets quite late. But when I was here in December, it was very, very cold. There was snow. But again, I'm a West Coaster, so I've been on both coasts, and I think I'm just naturally more inclined to say the West Coast is the best coast. Sorry for the people in New York. My co-founder's in New York, so.

Really?

Yep.

That's funny. So actually walk us through, I mean, kind of for the audience, you know, who may not know much about sort of Fireflies, or, you know, you yourself, can you give us a quick round of introductions to both you, your background, and then Fireflies?

Sure. I'll start with Fireflies. It's an AI meeting assistant that joins your Zoom or Google Meet or video— other video conferencing calls, and it takes notes, it summarizes the conversations, it analyzes the conversations, and helps you remember everything you've ever discussed, ever. So That is in a nutshell what Fireflies does. It's that meeting assistant that follows you around. It's that note-taker bot on your Zoom calls or Google Meet calls or Microsoft Teams calls. So that's what we do. And I think the inspiration behind it or how we kind of got started was really around we spend so much of our time in meetings. And, you know, if you're a knowledge worker, you're spending more than 50, 60% of your time and meetings are very expensive and you tend to repeat yourself. That was the inspiration behind it. Happy to go a little bit more into detail. And then my background is I graduated from UPenn and then I went and worked as a PM at Microsoft in their collaboration suite. We worked on a whole host of different tools. I was actually on their first growth team at Office and very much a data guy, A/B experimentation, those sort of things. I was working on that. And the reason I left Microsoft was I had a truly amazing time there. I was actually supposed to go to grad school, but then I had this summer break before grad school started. And that's when I went and met my co-founder in Boston. And we were just tinkering away at a couple different projects. And then I thought, hey, this doesn't feel like work. This is more fun. I just wanna do this instead of grad school. So I ended up choosing this route.

Yeah, there must have been a, You know, pretty big leap for you going from a, you know, remarkably stable and lucrative job at Microsoft to, you know, the founder path, right, which can be a little rocky. I know a lot of the listeners, a lot of my audience are aspiring founders, right? They're like, they have these beautiful golden handcuffs at either like Goldman Sachs or, you know, the Big 3 or, you know, tech companies. What was that journey like for you? And like, what gave you that— spark that made you finally take the leap?

So one thing for sure is being at a larger company, you naturally have more stability. You have very defined roles and responsibilities. The benefits and the salary is unparalleled, especially some of these, you know, big tech companies, FAANG, or, you know, the Facebooks and Amazons of the world, the Microsofts of the world, who pay you really well. And you are focused on a very small subset of problems. Also for someone that's just coming out of college that wants structure and established practices, it works great. Like it's a really good starting point for people that just wanna get off the ground. And so that work-life balance, like it's so nice to be able to turn off your laptop after 5:00 PM, right? At a startup, none of those exist. So I'm almost like, I don't wanna discourage anyone, but it's like when you're getting in, you have to be so passionate about working on things Putting in crazy hours, waking up at 3:00 AM to deal with server outages. Those are the sort of things that you have to deal with. It is not as glamorous as Silicon Valley makes it out to be, or TV shows make it out to be. But you have to be so in love with the problem that you're solving and what you are trying to do and the team that you're working with that it supersedes all of that. And I think my motivation was just from my personality type. I always wanted to build things and I wanted to build things fast and I wanted to have a sense of ownership over how it's shaped up. And so usually at a larger company, you have to earn your way up to do that. Like before you can own a product, you have to first own a feature. Before you can own a feature, you have to maybe work on some small incremental bugs. So I wanted to kind of jump that and go straight to building and owning a product end to end. So that was the inspiration. And I always felt that at startups, 1 year is equivalent to 5 years. So you can get a lot more experience and do a lot of things there. And I felt like experience was more important for me at that point in my journey, being a 20-something out of college versus optimizing for that pay. You know, I think even after I've left, I've kind of seen with a lot of my friends, that have joined much later, the pay is astronomical, especially 2020, '21. It was incredible, right? So in my mind, I'm like, wow, like, why would anyone want to go to a startup and leave when you're getting paid such a high salary, right? So yeah, I mean, it's the motivation and the desire to build that really led me in this path.

I've seen that as a pattern, like, among a lot of the founders that I interview, it's a desire to build and shape you know, what their vision is or idea is that I think drives them to sleepless nights and, you know, to take very emotional sort of leaps into entrepreneurship, because it's really logically, like to your point earlier, it is the— it is not the smartest financial decision, right, to be an entrepreneur in the early days at least, but in the long scheme of things, of course, it is the best thing you can do, but oftentimes that reward is far out. and you may or may not get it, especially if you're on the venture path like you are. It's a 10-year game, as we all— as we know. It's not a bootstrapped, you know, month over month, like 3, 4 months you can get to profitability game, right? So, um, it has to be something that's deeply moving and some connected to your calling. Like, you just feel like that has to be— there has to be some irrational urge to build. And for you, Krish, I've seen from your profile It seems like you've always tinkered with projects in college, during your internships and after. How would you describe sort of when you learned that, man, I'm meant to be a founder? Like, when was it crystal clear that, okay, I'm meant to be a founder?

It's funny you say that because when I was going through high school, I thought I was meant to be a doctor. And that's how I started my college journey. Didn't really think I was going to go into tech or—

Wait, you didn't go to med school?

I didn't go to med school, but I started off as a pre-med. I was shadowing— Yeah.

Wow.

Like doctors. I was doing research neuroscience stuff at Stanford. Everything I was doing that first, like that summer leading up to college and the first semester of college was around all of these things. I had done like 2 cognitive science research stuff, like worked on like certain papers that were being published. So it was a very different track. And it was only in like that first year of college, what really turned things around for me was one, I realized I didn't wanna do this for another 15 years and just study. Yeah. I had already done that in high school and, you know, I think I was a relatively good student in school, but I wanted to be more, I wouldn't say entrepreneurial, but I wanted to get into doing things that were anything other than medicine. 'Cause I just didn't wanna study and I wanted to do actual things. And so as I was like writing down my resume, I'm like, wow, I have no tangible skills other than reading papers and research and writing stuff. This isn't very valuable in the real world. So that's when I started working on several— I just started applying to a lot of different clubs. I explored a lot of different fields. And I think that's an amalgamation of what all of all the things that I've done. So completely unrelated, but I remember that there was an opportunity where the, there was this tax board that was being created at our university and they needed to look back at, hey, do we need to increase tuitions or how does tuition get allocated or the money that we have get allocated to all the different groups on campus? So I kind of went in there and had to learn a lot of these things. You run the math and you say, hey, let's do what's best for the students. Let's not raise. And instead, here's where we can cut costs and be very effective. So I always felt like understanding your customer or understanding who you're trying to solve problems for was really the starting point for me. And that was a very non-technical role. It was very business, but very non-technical role. After that, I spent a summer learning to code, especially my freshman summer, worked on a project with a friend where we didn't really have internships. So we said, hey, let's make our own internship and let's work on something. classic thing that a lot of freshmen do. And so we ended up building pre-Facebook events, like a tool that would let you know all the activities that were going on around campus. And then we got a bunch of college ambassadors. We got a couple thousand students to start using it across almost 2 dozen universities. And it was so fun putting things together, starting something from scratch, and then having people to work on it with you, right? You're like creating something out of nothing. And that gave me the thrill of entrepreneurship. And once I went to my sophomore year at Penn, I started working with a couple MBAs who were doing a project out of the Wharton Venture Initiation Program. So Wharton will fund a bunch of startups. And so I worked there. I was like, really, after the 2 founders, it was just me and you're wearing multiple hats doing a bunch of different things. And through that process, I felt like, you know, if I go and apply for a traditional software engineering role, Am I gonna have as much fun as this? So that's when I kind of got into the PM side of stuff. I started working on a bunch of different things. Again, you wear a lot of different hats, but it actually felt easier to me than working at like an early stage, you know, 3, 4 person company. So that thrill, like, I guess you're a thrill junkie, like you want that craziness. Right. That kind of led me to the route and, you know, ended up meeting my co-founder that way too, where we worked on hackathons together. And worked on a lot of different projects.

Who did you meet? Was he at Microsoft? Where did you meet him?

So I met him while actually in college. And so I went to Penn, he went to MIT. We met through mutual friends. We worked on—

What's his name? I'm gonna show up.

Yeah, so Sam Udetong, my co-founder and CTO. We've been, yeah, we've been at different projects since 2015. So we knew each other several years before we formally started Fireflies. Yeah.

So, Fireflies began when? When did it begin? 2019?

So, the product that we launched, what happened in 2020, and then before that, 2018 was when we started building out the core voice technology, 2019. Prior to that, we were doing a whole bunch of things like consulting, making ends meet, bootstrapping. You know, my co-founder has a whole post on moving to San Francisco. Wait, when did you leave Microsoft? I left Microsoft 2016, end of 2016. Yeah.

So from almost that end of 2016 to eventually when Fireflies took off, you were kind of like, kind of making it, you know, making it happen, like surviving.

Well, it's, uh, you— we started off with one idea. It's pivoted into like 5 or 6 things before we landed on, uh, the voice assistant, the meeting assistant that, you know, today. Uh, but that's what I say, like those 2 years from 2017 to 2019, was like walking through a desert without a map and realizing, wow, is this really worth it? Every day

you're

soul-searching and realizing,

you know what, I'm a bigger fan of you because of those 2 years than all of the rest of the success. Because you kind of, you know, that's the hard part, right? Most people would quit and go back to either Microsoft— anybody would have taken you.

It's tough because you're gonna get different looks from different people, your own family, right? Everything else.

But you

have to, you have to,

uh, How did you survive the desi, you know, family pressure, man? That's like hard. That's the harder thing.

Yeah, I would definitely say that I'm very independent in terms of making my own decisions. So like my parents were surprised when I got into a good college or when I got into Microsoft or when I do this. So I always, I think, surprised them or told them last minute. And so even till I was leaving Microsoft, they were under the impression that I was going to go to Cambridge for grad school. And it was only like 2 weeks before I have to fly out to London that I said, yeah, I'm not going to do that. I'm gonna go work on a startup.

So your best defense mechanism is just delaying until the very tippy top. Yeah,

probably not good relationship advice if you're in a relationship or you're married, but in general, that's how I kind of operate, was like, I don't want to talk about it until I know for sure, right? So that was kind of the thing. It was very surprising. I think my parents were definitely patient the first year, but then second year, you know, they'd be like, yeah, you know, if you were working at Microsoft, you would be settled down now. You could buy a house, you could do these things. So, it's the common pressure of being in an Indian household. Right. So, I think 2018 was really, really tough because that's where, you know, pedal to the metal, that's where it really matters. I remember making a trip out to India in 2018. We were hiring, we were basically doing whatever we can to keep like our operating costs low and just try to survive. And then in 2019, things started to change. We solved some core things around the technology. We started having customers, we started building out a product. And then that's when we realized, hey, this is something that we can actually raise money around. Because I actually didn't wanna raise, despite all that desperation and craziness, we took some money from angel investors, but I didn't wanna go out and raise a big institutional round without actually building a product and being confident in it. Because I wanted to de-risk the technology for myself.

Yeah.

Whereas, you know, 2021, 2022, you could walk in and say, hey, I'm so-and-so from Google or Microsoft, give me like $5 million, I'm gonna go build something. Yeah. If you have some reputation, like they were basically throwing money at anyone.

Yeah, exactly.

Very different times when we were starting, but I think that discipline helped.

And also the 2 things that stand out, one, this conscious choice of not trying to raise before you felt conviction, but the other thing that stands out is that, like we were talking about this at the beginning of this, sort of when we were interacting earlier is that Fireflies and the way you were leveraging this NLP was pre-generative AI hype cycle that began, I think really, you know, mid last year maybe. Yeah. I think with the OpenAI ChatGPT. So you were doing this in 2019, 2020.

2019. 2020, 2021. The idea of sequence-to-sequence modeling or to be able to detect action items without any generative AI helping you or LLMs helping you. to be able to detect filler words inside a conversation. Yeah. Dates and next steps. This is old school traditional NLP that we had to do. It's very hard. And today the OpenAI API essentially makes it so easy that I think anyone that can write a few lines of code or even use Zapier can do crazy ML that maybe 5 years ago only PhDs could do. at high levels of accuracy. Wow. And you have to train models, fine-tune them, make sure they're not overfitted, and it changes for different datasets. This also influenced a lot of companies' ability to build products. So because AI wasn't as good back then, people took very niche verticals, very focused areas like, oh, I'm gonna do this, but only for call centers, or I'm gonna only do this, but only do it for sales conversations. Our belief was we want to build this for every type of use case. That means there's going to be a lot of different types of meetings, a lot of different type of conversational data. But our bet was the technology will eventually be good enough where you can have general models that support everything. And I think that bet paid off. But when we were starting, the feedback and advice we got from investors, VCs, other folks Advisors was you need to pick a vertical that's going to pay you a ton of money. You need to do enterprise sales and you need to be very niche and you need to go top down. That was the flavor of the month. Like we had during 2020-21, product-led growth, PLG, all of these things become the hype cycle because companies like Slack and Zoom took off. But when we were starting, we were doing PLG, people discouraged it, right? And we were building a horizontal product or we were building AI/ML. They're like, This is an unknown space, so we don't want to invest in this area. So yeah, I think it feels nice to be vindicated, but even like the way we built our company, they're like, what do you mean you're remote from day one? Like you have to have everyone in Silicon Valley, everyone in the office here. So we didn't follow any of the conventional rules. We were distributed, we were horizontal, we were product-led growth. And so not sales-led, but really around, you know, who charges $10 per customer when you can go upmarket and charge like $150 per seat? Uh, we did all like the crazy things, uh, but it was a core belief that AI can be democratized and that AI will get better.

Right. That's brilliant. Did you see the inflection point of OpenAI and generative AI? Did you sort of see that coming?

I did not. think that it would be as good as it turned out to be. We had almost a 1-year head start with working with OpenAI and getting early access to GPT-3. I think for a lot of companies too, the inflection point was when they saw ChatGPT come to fruition. And so that was really like the game-changing point for a lot of companies where, wow.

What about it really blew your mind, do you think? Like if you look back, Like, what was that sort of paradigm shift for you?

The idea that you can do single-shot or zero-shot learning where you can give it data and have it classify it or answer things, it was definitely, that was the game changer. You don't need like millions of data points. Anyone can implement AI with a few lines of code, right? So that was something that was super powerful. What I do believe was all the companies that were already doing work pre-ChatGPT, pre-generative AI, if they were moving fast enough, they can take full advantage of this wave. There's a lot of companies that are now spinning up where, oh, I see generative AI as an opportunity. I'm gonna build a startup on top of this generative AI model. We didn't do that. So like it was, it's different. So I think you're gonna see a lot of shiny objects as a result where people are gonna see an opportunity and try to do it. But the companies that have already been doing the traditional SaaS stuff already have the users, already have the data, already have built out like a process. It's almost a catalyst. You're adding fuel to the fire rather than using that as like the way to get started. So we were in like a unique position, like where we were very lucky on timing. We launched officially in January 2020. Then the pandemic happened, everyone went remote and

What is— which is a great, again, paradigm shift, right?

For everyone wanting an AI note-taker 'cause they're having Zoom exhaustion, too many meetings. So it was a great starting point to get off.

You got lucky twice, you know?

Yeah, so that definitely, like the inflection point took off there. There was a technology inflection point between 2018 to 2020, really around 2019 where the cost of transcription went down, the quality went up. Yeah, yeah. And you have to build—

Dude, what happened there? I remember 5 years ago if somebody came to me for this exact podcast or any kind of like, you know, audio thing that I would be doing and they would talk about per minute. And I thought that was like, that was so weird. And I was like, why would somebody pay anything per minute? You know, but that was the standard. What happened? I'm curious technology-wise. Yeah. What happened there? Because now anything in terms of you guys or any other transcription is just like, doesn't even count the minutes. It's only mostly per month, $10 or $12. What happened there?

So I think that even today, if you want to go get a human transcriber to transcribe an hour-long podcast, it will cost you about $1.50 to $2. Maybe at best you'll get $1 per minute for, you know, the podcast. So a 60-minute podcast could cost anywhere from $60 to $100, right? And with Fireflies, like the $10 per month subscription, you are essentially for $120 for the year, you are getting almost like unlimited meetings being transcribed, right? It's not even comparable. One meeting versus one year's worth of data. So there's a tectonic shift there. One is automatic speech recognition got fundamentally better to the point where you can't really distinguish between a human Versus like the machine generating it. The customization of unique vernacular got better as well. So it was picking up things.

Wait, so when you say this automatic speech recognition, are you talking about the chip level got better or are you talking about the software service layer level?

Both was important because you need tons of data to train on these things. You need to have access to GPUs, you need to be able to actually train on tons and tons of data and Also, people took different approaches, like the new conformer approaches that were coming in. So they were saying, hey, maybe we can do this a different way. We can tackle this problem a different way. One where, yes, it's a model improvement. It's also like the amount of quality of data that was available. Once a few companies are able to show you that this is possible, there's a lot of other folks that are starting to work on this stuff, right? And I think we might in the near future, we already have some open source ASR stuff, it's one thing to use it and run it and train the models, but in the long run, it's going to be like the cost of compute, just like storage. When storage first came out, when Dropbox first came out, 1 gig was impressive. When Yahoo Mail came out, you had 25 megabytes of storage, and people thought that was like, wow, that's a lot, right? And you have to delete your inbox constantly. And then Gmail comes outta nowhere and says, we're gonna give you 1 gig for free, no questions asked. Right. And all of a sudden you can go back to emails that you've sent 2, 3, 4 years ago and search through them. In the same way, Fireflies has come in and said, hey, we'll help you remember every conversation you've ever had. So I can search back through a conversation I had 2 years ago with an important vendor or a customer and have perfect recall. So you're changing the dynamics of the game just through the pure Computational storage architecture and improvement level. So cost is a huge driving factor for enabling this. And the thing is, we started when that was not clear. When people said, "I'm not so sure I can bet on costs ever becoming cheaper." So you have to be a little crazy to start early. But if you are right and you take that bet, it pays off, right? Having some of that early advantage to get going. Yeah. So that's why, like I think even 2019, 2018, 2019. 2020 was really tough. We're like, how good can this technology get? And then how cheap can it get to the point where, you know, this is no longer defensible? Like, this is not like something where we have to worry about. So yeah, it was a huge thing.

What were some of the factors that helped you build conviction early on?

I think for us, we started looking at doing things 2 different ways. We first A/B experimented with humans in the loop, right? Having a small set of customers. Like, we know that people are willing to pay if it's perfect, but the technology itself was not perfect. So are they willing to pay less? Are they not gonna willing to use it at all? So you have to figure out that satisfaction curve where people are willing to say, okay, I'm going to try and use this. So once we got real customer feedback and validation that the automated approach can work. Like, yes, you're gonna have churn in the early days, and yes, people are not gonna be happy, but you have to just keep chipping away at the problem. And the challenge with Silicon Valley is that it's not about faking it till you make it. It's about making it so you don't have to fake it. And it's a challenge. It's a struggle. So a lot of times when you're going through and seeing all of these other companies, you know, perception can always be mistaken as reality. So these companies seem like they're doing more than they actually do. So even when some of— there's been so many competitors that have come and then died, where they promised things to do that they couldn't really justify, or they were doing it in a very expensive way. So we've seen almost a dozen competitors come up and die since 2018 in this space. So I'm very, very practical and I'm trying to be as pragmatic as possible. I don't let the glitz of the Valley like overshadow all the things that are happening because at the end of the day, no matter who's working on a startup, it's a hall of mirrors and you don't wanna let all of these illusions of someone is doing something better that we don't have like ruin you or change you from the focus that you have, right? Think about Amazon when it first started versus them shipping from, you know, the garage books to now being able to ship anything within a matter of a day, sometimes even same-day delivery. So you have to keep iterating, chipping away at the problem. And that's when you look back after a couple years and said, wow, we can do a year's worth of data. What used to traditionally cost 1 hour, we can do it for an entire year. And that's just constant iteration. That, that, that is a belief that I didn't even have till maybe a year ago. Um, but that's just a function of how fast the technology is evolving and improving.

Yeah. So if you were to give one piece of advice for aspiring founders in terms of idea selection, right, picking the right problem to work on, what would, what would that be? I mean, one, one piece of advice, but you can break it down to 2 or 3 pieces.

Yeah, though there's a lot of different ways to think about this and how you want to consider it. So one thing that I will say is that product-market fit that everyone keeps talking about, like that, that promised land of you know it when you have it, it doesn't happen overnight. And it's not one thing that will lead to it. It is a sequence of iterations and trial and error. that will lead to it. So no matter whether you're a first-time founder or a second-time founder, the assumption that you start with doesn't necessarily mean that's what the end product is gonna be. You have to keep iterating and it might be on that 5th or 6th iteration and slight improvements. You're just steering the ship bit by bit, bit by bit in the right direction. It's not like a full 90-degree turn. A lot of those stories that you hear, the narratives that you hear from people about completely pivoting to something else rarely works. And it's also really a stroke of luck if it does work for them. So it is a journey where, you know, you have to constantly iterate and ship. And then the other thing is 95% of startups fail. So failure at a startup, like in life, is only final if you stop trying. So as maybe that sounds a little cliché, but really, You're gonna fall a lot and you need to be okay with imperfections. That was something really hard for me, like where I wanted to ship the most beautiful product on day 1, something that was perfect. But you actually are sometimes embarrassed by, oh wow, that's how we used to be like. You need to be. Because if you look back a year and say if you're not embarrassed— so if you look back a year and if you're not embarrassed at the product that you shipped and how much better it is now than a year ago, then you're doing something wrong. That's how you mark progress. Like even not just products, but yourself as a person. Who were you a year ago and how are you today? And if you can't like say, I am significantly better as a person, like in terms of fitness, in terms of health, in terms of like knowledge, then you're not moving in the right direction, right? So it's the same thing applies to startups.

Yeah. Reminds me of this podcast itself. I always joke the first episode, it looks like I shot this in a dark alley. Like the lighting is bad, everything is bad, you know, but this is the 60— I think we're on the 63rd or 64th episode, and when I look back at some of the stuff, I'm like, man, I was such a rookie. That's the feeling you want, right? You want to feel like, man, that we were such, you know, noobs at this, right, at the time. So no, that's a great answer. So looking at markets and what VCs look for, if you had to sort of, I mean, I guess give your younger version Right, the younger Krish from 2018 about fundraising, what would be some of the advice that you would give that Krish given this knowledge that you have today?

So when we raise money versus what happened between 2021 to 2022 to where we are at now, so I think we have reverted back to the mean. It's much harder to raise money. The terms are not as nice. So there is a lot higher of a barrier in order to raise money now. I'm sure like if you're an AI company, you can go and create some sort of magic out there in the market. But in reality, people were raising money just based on a deck and based on their work experience 2 years ago. And today the question is, do you have customers? Do you have a product? And do you have revenue? When we had started, it was a little bit— easier than maybe what it was today. Still difficult where do you have a product, especially if you have no background? Do you have customers? And do you have like an opportunity to start making revenue? So it really depends on that part. It just depends on who. But I think today VCs are a lot more focused on is this gonna be a disciplined company? There's gonna be some folks that bet on— Yeah. Like this big future and people can sell a big narrative. I always hesitate on that because if you're raising hundreds of millions of dollars and not able to live up to that expectation, you didn't do right by your customers. You didn't do right by your employees. If you get down rounds, you didn't do right by your VCs as well. So I take a different approach where I say like try to be as scrappy as possible. Build something where you're removing as many layers of risk so that when someone looks at it, they say, oh wow, I actually don't see as many risks as I would for other companies. So the VC needs to feel like you are a diamond in the rough. And the only way you can do that is by showing points of success and improving. So a lot of times when you're fundraising, you might first meet them early in the journey, and they might not be ready to invest at that point in time, but show the progress over time. Like, hey, you know, when we first met, this was just an idea. Here we built a prototype. Here we're talking to a customer. Here We have a pilot contract in hand here, like we're doing this. So always show progress in the right direction. That always impresses early stage VCs. 'Cause at the end of the day, the early stage VCs are betting more on you as founders and your work ethic than your idea, because the idea might change, you might pivot, but the ability to show continuous execution. is really, really the most important signal if you don't have already an established brand or name or reputation. Because to be honest, when my co-founder and I moved to Silicon Valley, we had some good names from a university point of view, but we didn't have any connections. We didn't have any network. We didn't know any like VCs off the bat. So it was something we had to like grind and prove. And again, there's a lot of noise because back then there were so many startups vying for attention. Yeah. So yeah, it was a grind.

Tell us about your fundraising rounds. Did you raise 2 rounds, 3, or any notable funds?

For our fundraising, we did a seed round, which was about $5 million in 2019, end of 2019.

Well, that must— wow, that 2019 $5 million is different from 2021.

Yeah, 2021 $5 million is like $15 million, right? It's nothing. Yeah, but it was a nice sizable round from Canaan Partners. And then—

oh, wow, that's a big name. Yeah.

Thank you. And then we raised our Series A from Coastal Ventures this past year. So that was— they did our Series A, and then our seed was done by Canaan. And before that, we had a lot of really awesome angel investors. So we didn't— we had like some small checks from different people. And then during our seed, we brought in some really awesome People like the head of product, the chief product officer at Slack, the CMO at Salesforce. So a lot of these people that have come in early in that journey. So some well-reputed names, the first engineer at Dropbox. So like a lot of like great names. Peter Reinhardt, who is the CEO of Segment, he sold his company to Twilio. So he came on board. Yeah. So there were a lot of really awesome names that came on during that seed. Who we've been able to lean on for advice and guidance and very, very happy about that. But yeah, it was fundraising was something where I don't like to oversell. And even my investors today tell me like, you know, you can be a little bit more aggressive in like what you're saying is like, no, I want to promise less and deliver more. Complete opposite of what you're supposed to do as a founder, right? You're supposed to sell the dream. Um, but I believe that it's more important to make reality happen, uh, and show that a dream is possible. So, it's the other way around, right? So,

um— Yeah, I love that. I mean, I think it's a— it's, uh, to be a founder, you have to be a little crazy, right? So, it's not that you're trading crazy and optimism to pessimism, cynicism. You don't have— I— what I'm— even just from the last 40 minutes with you, you're stubbornly optimistic, like, the fact that you, you know, made it through that desert storm like you talked about, but you're pragmatic. And I think that's a great quality. You know, you wanna have that optimism about a great future, but also live in the reality. Like, where are we? What are we? What can we do technology-wise?

So Silicon Valley is a land of unicorns, but we have to remember that it's the workhorses that often go the distance. Yeah.

Yeah. Wow. So Krish, one last question. This is because, you know, this is the Building Public podcast. So of course we have to talk about community and building with them. How much of a role did your early community play in your story? You know, early users, like you said, early angels, all of those people.

I think the one best advice that comes out of like Y Combinator or anywhere else that people may wanna do more seriously than ever is just listen to your customers. Because more often than not, you'll know exactly what it is that is required. And that to me was the community of beta users, early testers, because the way we launched Fireflies was actually through our network of friends that started using it. And then the virality of seeing the note-taker in meetings and the word of mouth helped it spread. But to this day, I look at every user that churns, they leave a piece of feedback. We ask for it. And when they cancel their subscription, I read it. I read it every single day for the last 3 years, why someone chose to leave Fireflies. And likewise, I read the compliments. I don't really care about the compliments. I don't take it to heart as much as I take the negative feedback. That keeps you grounded. So imagine waking up reading, you know, over the span of a year, hundreds of pieces of feedback that tell you why you're not good, right? Or why you were not good enough. And you literally use that to motivate you and drive you. So I think that's, that's actually been a really important thing for us. But that community that you end up having with people that are giving you feedback, it is the ultimate source of truth. It's the ultimate knowledge. Same thing when you're doing sales or when you're doing partnerships or when you have a community of folks that are reviewing and helping build out your roadmap. So it's those people that have skin in the game that are gonna always give you their genuine feedback, because why else would they pay for the product, right? So I always like to kind of keep that as like the most candid thing.

You know what, it just reminds me of one of my friends, this guy I brought on the podcast a couple episodes ago, his name's Lewis. He uses this phrase that I've been using myself as well, where often people don't know that you're looking for feedback. I mean, and they may think they're looking for compliments, right? And so they may just give you, they may just like shower you with nice things and compliments about the product or whatever. And so he uses the phrase where he said, I'm hungry for constructive criticism or constructive feedback. Go at it. Like, feel free. So basically the language, I'm paraphrasing obviously, But it kind of opens up this candid two-way door, you know, with your customers and allows them to realize that, oh, right now is the time for me to be open about where's the room for improvement, right? If you don't say that out explicitly, often you just end up getting just nice compliments, which is good, but it's not, you know, it's not directionally helping you with the roadmap, right? I

think it's, again, the perception can often be mistaken as reality. So if you are in an echo chamber where, you know, whether you're a politician, whether you're a founder, whether you're the CEO, if all you hear are nice things or good things, you forget to know what is going on. Yeah. And I'm a, I was someone that I loved history and mythology and like learning about like kingdoms and empires of the past. And I used to constantly hear that, like, you know, like a really good king back in the days, would actually go in disguise and go figure out like what's going on, what are the actual problems that people are having versus having his lieutenants like tell him while he sits on a throne and, you know, is laid back and having a chill life. So you have to put yourself in those like hard situations as an engineer, as a product person, as a founder to learn that stuff. And I think that people say don't take it personally. I definitely take it very personally, right? Because when you have 97 out of 100 people saying good things about you, I'm looking for those three detractors why they weren't happy, and that's the stuff that keeps me up at night. Probably not healthy as like as a founder, but like those are the things that I take really personally. There were times when I wouldn't even eat lunch because I used to do a lot of support tickets when we were a small team, and like I wouldn't eat lunch until that person's problem was like solved, or if they said something about like oh you know your product it just doesn't work you know it's just not like It's not good. And then I would think about it like a lot. So to the point where I didn't want someone— if someone said something bad about the product, it's a negative reflection on me. So almost as if they were insulting me. So that's why I took it personally. Yeah.

But that was a fun answer. So Krish, last question is, what's next for you and Fireflies? Like in terms of, you know, business roadmap-wise, and then personally, what are you looking forward to? You know, as a founder? So

for Fireflies, we found that what we're doing with the meeting summaries is just the tip of the iceberg. We've built this end-to-end platform that is capturing the most important knowledge and information for an organization. And we are very excited about a sequence of new products that we will release. Very similarly, what we're calling apps. The idea is that we wanna allow users to extract any type of information they want from their meetings. So if you had a recruiting call with a candidate, you can have an app that will automatically score the candidate. If you have a sales call, you can have it pull out— usually salespeople have like these acronyms called BANT, like Budget, Authority, Next Step, Timeline, or MEDIC, some sort of sales framework, right? Automatically pull that out and put it into Salesforce. Uh, so a lot of really cool things, maybe automatically creating a lead scoring app on top of conversations. So you have this really precious, valuable data. How can we allow you to build apps on top of it, low-code or no-code? So really, the summaries is just gonna be one app on top of Fireflies. There's gonna be hundreds of other apps. It's similar to when Slack released the Slack bots, right, for their ecosystem and their app store. Wow. We're very excited about like this app store that we're building. And all of it is gonna be powered by generative AI. We're working very closely with OpenAI on some of this stuff. We've released 2 apps already. One is called Super Summaries, which will take your regular summary and just blow it up like to the next level where it'll give you a paragraph-level summary. It'll give you action items. It'll give you an outline of what was talked about. You can click on different parts. It'll give you action items. And then we released an app called Ask Fred where You don't wanna go listen to the entire call or read the transcript or even the notes. Just ask Fred like questions about, you know, what was KP's most transformational moment on his entrepreneurial journey? And it will pull that out, right? All that information without you having. So think about it as like natural, like how you ask Alexa or Siri questions. We built Ask Fred. So we're gonna have many more apps over the span of time.

And what do you feed to that? You feed just the— the previous meeting or whatever that meeting that Firefly's joined?

Yeah, so today it'll be based on the context, contents of the meeting, and you get to define like what you want to pull out of it. So it's almost like taking all the folks that want to like get certain specific things out of it. So for example, if someone says, I want my board meetings outputs to look like this. Okay, great. You can go customize. You can just literally go tell the AI saying, Create notes based on this outline or this structure, or on a sales call. I want to know what their next steps are and how long it's going to take to implement it. And after every sales call, if I get that information pushed to Slack, I don't even have to sit and review the call, right? I know like, oh, they want to buy ten seats. They need to implement this by next Friday. This is their blockers. Automatically, like you are able to understand a conversation. So everyone has different needs and wants different things. The summaries was just one of those needs, but there's hundreds of other needs we can satisfy.

It's also amazing how long you stood with summaries too. You didn't like simply like, you know, move on quickly. And now you, I feel like you've earned the right to move on from like, not just move on, but like append to the summaries as a feature and the other things you're talking about. Right. 'Cause for the longest time, that's all you did, right?

Summaries is very subjective because there is no perfect summary. Everyone's going to want something different. Yeah. So I think we're at a point now where I can have the summaries be written like a human, and it's almost indistinguishable that a human wrote it. Again, there's gonna be sometimes where the summary's really good, sometimes where it's not like that good, but we are controlling that variability to the best of our ability. And I think the transcription summaries is why you'll come to Fireflies, but then you will stay for the knowledge base that That is being created and all the apps and workflows that you can do on top of it.

So come for the summaries, stay for the knowledge base and apps.

Yeah. And let Fireflies turn all the conversations like voice into actions. That's the old, that was the dream, right? I say something on a meeting like, hey, we need to send this contract out to so-and-so before next Friday. This is what they need in that contract. Imagine a future where Fireflies is automatically updating that contract, generating and say, hey, this is the contract you want to send, right? Uh, go

ahead. Yeah, that's just wild. That's wild, man. Awesome. You know, this has been such a blast, Krish. Enjoyed getting to chat with you. I would love to go try the app right away. I mean, I, I kind of know the website, like I was saying at the beginning. Um, I saw the transformation from what I saw, I don't know, 3 years ago to now with the generative AI. But given what I heard from your description, some of the cool things like Super Summaries and Ask Fred, I can't wait to go play with some of these things and give you some feedback. Awesome. Thanks for being on the show and have a great rest of your day.

I had a lot of fun. Thanks. Thanks, KP.