YC Root AccessRebuilding Customer Support for the AI Era
EVERY SPOKEN WORD
75 min read · 14,773 words- 0:00 – 0:05
Intro
- HTHarj Taggar
[upbeat music]
- 0:05 – 1:56
What Pylon is today: AI-native B2B customer support platform (scale, customers, funding)
- HTHarj Taggar
Hi, everyone. Uh, I'm excited to be joined here today by Marty, Robert, and Advith, the co-founders of Pylon. Pylon went through YC in the Winter '23 batch, and is one of our fastest-growing companies over the last few years. Uh, I think what's especially exciting about their story is they took a while to find the idea and pivoted through several ideas, so we're gonna chat about that today. Uh, guys, thanks so much for joining us.
- ADAdvith
Thanks, Harj.
- HTHarj Taggar
Why don't we kick this off by just telling us, what does Pylon do? What's the problem it solves?
- MAMarty
Yeah. So Pylon is an AI customer support platform specifically built for B2B companies. We replace products like Zendesk, Intercom, Salesforce Service Cloud, but very AI-native and built for a B2B persona.
- HTHarj Taggar
Could you give us a, uh... Share anything, um, to give us a sense of the scale that Pylon operates at today?
- MAMarty
Yeah, so I think we're 79 people as of this week. Um, we're three years old, uh, three years since, uh, we spoke to and met you, Harj, um-
- HTHarj Taggar
The Winter '23 YC batch.
- MAMarty
Winter '23, and yeah, I mean, back then, it was just the three of us. Um, we actually pivoted through many ideas, as, as we'll talk about. But, uh, yeah, 79 people today, three years old, um, eight digits in ARR, uh, and last year we grew 5.35X.
- ADAdvith
Over 1,000 customers now.
- HTHarj Taggar
Nice. Can you mention any of the biggest customers you have?
- ADAdvith
Yeah, folks like Honeycomb, Linear, Incident IO, Applied Intuition-
- MAMarty
Deel. Yeah.
- HTHarj Taggar
And how many rounds of funding have you raised?
- MAMarty
Um, so we're Series B now. We've raised 51 million total from, of course, YC, General Catalyst for our seed, Andreessen for our Series A, and Bain Capital Ventures for B.
- HTHarj Taggar
Cool. I wanna come back to sort of Pylon as it is today in a little bit, but I actually think what's really interesting about your story is it took you a while to arrive at this idea. Um, so why don't we kind of actually go all the way back, and maybe just tell about- tell us a little bit about each of you individually, so of your, your backgrounds, um, why you got interested in doing startups, and then maybe we can talk a little bit about how you started thinking about startup ideas.
- 1:56 – 4:41
Advith & Robert’s origin story: Caltech, hackathon ops, and startup pull
- MAMarty
Yeah. Advith, you wanna start?
- ADAdvith
Yeah, I can start. So I think... Uh, so Robert and I went to Caltech, and I think in college at least was when I first started getting interested in startups. I was considering being, like, a math major, and then when I went to college, I was like: "Oh, I guess I... Computer science is some applied version of math." And then met Robert freshman year doing, like, a puzzle competition. Um, we're both super into puzzles, and then we ended up running our school hackathon to- together for a couple of years, which ended up being basically like running a company, minus building a product. It's a lot of, like, organizing tons of people, raising money, just, like, holding an event, essentially. So like, building a team that does something, minus building a product. And so, uh, through that, uh, like, I guess, built the muscle of how to work together, with Robert especially, and then that-
- RORobert
Yeah. I think the funny thing is, we were, uh, work colleagues in college. Like, we weren't really friends. [chuckles]
- ADAdvith
Yeah, yeah. So we, like, had a great working relationship where we'd work on stuff. Like, Robert would take the night shift of, like, midnight to 8:00 a.m., and then I would take, like, the rest of the day shift, like answering emails and, like, doing all these, like, company logistics stuff. And then that summer of sophomore, Robert's sophomore year, he did this program called KP Fellows, which is this, like, entrepreneurial, like, internship program, and, uh, he was at DoorDash that summer, and I, I was interning in the Bay as well, but I just... That program has a lot of f- ex-fou- or, like, people who become founders later, and so got to see what was going on through that. And then the next summer, I did it, which is when I met Marty. I was interning at Slack. Marty was at Airbnb. And I guess the TLDR about why starting a company is just wanted to do it even in college, just didn't have a thing to do it about. And so, um, uh, like, a couple of years after, um, graduating, w- we were all at startups, and then I messaged Robert, like, right after the pandemic, being like: "I feel like I'm wasting my life." So my... It was on my 23rd birthday. It was like a existential crisis of a sort. Just felt like I wasn't really doing anything. I was getting bored, and messaged Robert and asked him-
- RORobert
That was on your birthday? [chuckles]
- ADAdvith
Yeah, yeah. We actually started our first meeting, like, on, on my birth- like, 6:00 a.m. on my birthday. Yeah.
- HTHarj Taggar
One thing I'm actually curious about, um, quickly is, so you both- you... Remind us, where were you doing the internships?
- ADAdvith
I was at Slack.
- RORobert
I was at DoorDash.
- HTHarj Taggar
Yeah, so those are both great companies. And Marty, you were-
- MAMarty
I was at Airbnb. Yeah.
- HTHarj Taggar
Airbnb. Okay, so all three of you were at, like, fantastic companies. I'm just curious, like, um, why did you decide to start your own thing versus joining one of those companies?
- ADAdvith
I, I, I just felt like at ev- every internship I was trying to work at a smaller and smaller company. I actually just felt like I was wasting my time, for lack of a better word. Like, my main... Not to insult Slack, 'cause our- a lot of our [chuckles] products today-
- HTHarj Taggar
[laughing]
- ADAdvith
... are based off of Slack. But at that scale, it was maybe a couple hundred people, and, like, my core internship project was they were doing a rename from, like, calling stuff teams to calling stuff workspaces, and most of my project was literally rename [chuckles] like renaming, and I just thought it was-
- HTHarj Taggar
Pre-AI.
- ADAdvith
I thought it was super waste of my time, and I actually quit. I, like, asked to leave the internship early 'cause I thought I was wasting my time. Yeah.
- 4:41 – 6:27
Marty’s path: from music dreams to founder mindset and leaving Airbnb
- MAMarty
Yeah, I, I think on, on my end at least, um... U- so in high school, I actually wanted to be a musician to start, so I had, like, written an album. I was in a band. Um, and then I saw The Social Network, and I saw Pirates of Silicon Valley. So I actually feel like movies made- like, just opened up the idea, like, some sort of, like, visualization of, like, what starting a company could look like, and the fact that anyone can kinda do it with a computer, um, with no resources. So I was already, like, personally interested in, uh, programming, had, like, watched some YouTube videos, got, like, JavaScript for Dummies, like, from a garage sale at some point, and so, you know, was doing that. Uh, and then watching those movies kind of made it, like, real that, like: "Oh, this could be much more than just, like, a fun little hobby." Yeah. And, and one way we've actually all described it is, I feel like at some point we just realized we couldn't not start a company. Like, you're just, like, sitting, doing whatever we were doing at other companies, and it's like, you just, in your gut, you already know the answer, which is you have to go try starting something, and, like, nothing is gonna make you happy outside of that. Like, you have to just go try it, even if it's a mistake.
- HTHarj Taggar
How about you, Robert? How did you know you wanted to-
- RORobert
Uh, yeah, I mean, I grew up in the Bay Area, so I've been surrounded by, like, tech startups my entire life. Um, I always imagined engineering as just like, like a hammer, and I- while I love hammering nails, I'd, I'd rather build a house than think about the house overall. Um, so that's always been my goal, and I think, like, every internship, I just kept going to startups or companies just to observe how people build companies and just, like, kind of gather up all the different ways people-... build a team. And I, the, and the, that's why the DoorDash experience was really cool, because I got to see, like, what a great, uh, company looks like at such an early stage. I really didn't start my journey, um, full throttle until Advith did reach out to me that one night. I was like, "Sure, let, let's do it." [chuckles]
- 6:27 – 10:20
Starting while employed: COVID remote advantage, ‘Grind Time,’ and first edtech mistake
- HTHarj Taggar
So yeah, just talk us through the story that, like, after you reached out to Robert, then talk us through what happened next, and how the three of you came together.
- ADAdvith
Actually, while I was- w- at that point when I was reaching out to Robert, I was living with Marty. So Marty and I were roommates after we graduated.
- RORobert
At that time.
- ADAdvith
Uh-
- RORobert
No.
- ADAdvith
No. Or we were-
- RORobert
It was during COVID.
- ADAdvith
It was during the pandemic, so we might have moved away for a bit-
- RORobert
Yeah
- ADAdvith
... but, like, yeah.
- RORobert
Yeah.
- ADAdvith
So Marty and I were roommates after we graduated, uh, for, like, a couple years, and Marty has w- had already, I think, taken the plunge before that, and, like, had started trying to do his own thing separately. At that point, I wasn't... I guess I wasn't ready to... I felt like Samsara, the company I was at, doing really well, so I wanted to just do it on the side for a bit. And so, messaged Robert, and, and then we started basically every morning... Or I guess it started at a weekly cadence, just, like, brainstorming. And then it very quickly turned into, like, user discovery, where we tried to turn to, like, the first set of people that, um, basically are good candidates for user discovery, based on, like, this book, The Mom Test. I think, like, YC's super all about too, and that was school administrators for us, 'cause we spent a bunch of time running our school hackathon. So we reached out to a bunch of the admins at school, just asking them what their problems were, and we landed on one problem that's not really a problem, [chuckles] but, uh, we thought it was a problem at the time for, like, the career center. The lady who runs the career center was someone we worked with a lot when we were in college.
- HTHarj Taggar
Mm-hmm.
- ADAdvith
And so yeah, it was basically, can we build a tool, like an alumni portal, so that students who have already graduated can reconnect with alumni for, like, career advice? And yeah-
- RORobert
And we made all the mistakes, uh, going through Pivot Hell. Like, we built it before really validating with anybody. We were trying to build something in edtech, which we know is, like, a probably a bad space to start out in. Um-
- HTHarj Taggar
And you guys were working full-time jobs at this point?
- ADAdvith
Correct.
- RORobert
Mm-hmm.
- HTHarj Taggar
Yeah, I mean, a question I feel like we get a lot at YC is, um, from people who want to start a company someday, is, "How do you do that? What does that look like day to day?" Like, you're working a full-time job, and you're trying to make progress on an idea.
- RORobert
Just... I think we were f- really fortunate because it was COVID, so that because everything was remote, we [chuckles] could actually take a lot of the calls-
- ADAdvith
There's a lot from home
- RORobert
... on the side. Yeah, yeah.
- ADAdvith
And I think at that point, at least I felt like I had gotten pretty good at my normal job, uh, if that makes sense, and so, like, could still be pretty effective without, like, putting in the same amount of time. And so we started with early morning calls. I remember at, like, 6:00 a.m. we would talk, and then at night, and then it was, like, on a weekly cadence to start, and then it became daily. And then we started filling in all of the empty slots within our days with essentially customer discovery calls.
- RORobert
Yeah, I remember my coworkers would make fun of us, 'cause we had a calendar event called Grind Time.
- ADAdvith
Oh, yeah. [chuckles]
- RORobert
And they'd be like, "What is that?" [laughing]
- HTHarj Taggar
[chuckles] Sure. And Grind Time is when you work on your startup ideas?
- RORobert
Oh, yeah, when... Yeah, yeah, yeah.
- HTHarj Taggar
Oh, yeah. Um, how about you, Marty? How were you, um, exploring ideas to them?
- MAMarty
Yeah, so I think a, a little before, um, you guys started messaging, working on stuff... And first off, I wanna call out, like, um, I feel like none of us were sitting still at any point. Like, I... Advith was making, like, a card game with someone else. Like, you built an escape room while we lived together in India. So I feel like even, like, while working at full-time jobs, I feel like all of us were probably doing something. Um, and so yeah, for me, I was at Airbnb for maybe, like, a year and a half, and I was just, like, kinda miserable. I was, like, looking around myself at everyone else, and I was like: Okay, do I wanna be an engineering manager? Do I wanna be a VP one day? Um, like, who around me would I trade places with? And the answer was no one. I just wasn't interested at all in switching roles with anyone who was ahead of me career-wise. And m- at the same time, I was looking at founders who were just kind of, like, pivoting around and, um, trying to build companies, and I still, like, thought that looked way more exciting. And I was like, "Damn, if I spend, you know, the next five years just working my way up at Airbnb, I'm- I already know I'm not gonna enjoy that outcome." At the same time, I'm looking at founders, even if they're not that successful, and just saying, like, "I'd switch places with them." So, you know, during Airbnb, I started doing, like, side projects with all sorts of people, many different sets of co-founders, um, and many different projects. Like, everything from, like, like, a dentist app thing to, uh, like, a 3D clothing manufacturing thing. Um, like, ev- anything I can think of.
- 10:20 – 13:25
Pivoting philosophy: refining validation loops and learning from dead ends
- HTHarj Taggar
Were there any differences in, like, your approach to testing out startup ideas early on versus the two of you?
- MAMarty
Um, I'm guessing that... We haven't talked about this explicitly.
- RORobert
No.
- MAMarty
Um, I think I, I can explain at least my philosophy, which was, like, find any problem [chuckles] and then just try to see if you can solve it. I think for me, well, one, you kind of, like, look at your own problems, and then, like, comb through that list. Uh, I feel like quickly you run out of things. Um, I feel like I didn't have that many problems that weren't, like, fun-oriented back then. Um, and then I started, like, just talking to people. And so for example, like, this, like, 3D, like, clothing printing thing was, uh, the... One of my friends who I met, um, in, in SF, he, like, h- had an idea related to that, so, uh, he came up with that. Then there was actually one of our ex-roommate's, um, ex-girlfriends was in dental school, and so, like, I started talking to her about, like, the, that dental thing. Um, and then eventually, at least for me, I realized I was only gonna make so much progress while working full-time at Airbnb. I, I wouldn't say I was, like, in a place where I was so comfortable that I could just, like, easily not work that much, and it was still in person. I think during the pandemic, like, you could just be at home.
- RORobert
Yeah.
- MAMarty
For me, I just realized I'm miserable at work. Um, I actually have enough money saved up to, like, take some risk, and then I... This was really helpful for me. I just, like, made a, like, best outcome, worst outcome doc, where I was just like, "Here's, like, the best thing that could happen if I leave Airbnb and try to start a company." And so, you know, wrote out, "Okay, it goes really well, like, make a bunch of money. Like, I love the job, and I no- I don't look back." The worst case is, like, "Okay, a couple of years of opportunity cost to being a software engineer. Um, you know, I, like... Maybe I struggle to find a job again," um, but, like, realistically, it didn't feel that bad. I felt like I would just get another software engineering job two years later. So yeah, I think writing it out helped a lot to understand, yeah, what, what is, like, the worst case, and then also then, um, pairing that with knowing I was gonna be unhappy for sure if I stayed.
- ADAdvith
... Yeah, I guess our process was, at least for finding ideas, it was kind of similar. It was, like, looking to the people closest that we felt like we could source problems from. I think we understood that we had [chuckles] no problems ourselves to, like, actually solve. And so, and luckily, we had been working for a few years, so there was... I think we made a long list of stuff related to what I was working on at Samsara, what Robert was working on at Affinity, and tried to, like, explore those threads. They ended up not leading to what felt like the biggest possible opportunity, but-
- RORobert
I think the cool thing is, like, of all those ideas that we had looked through, they do end up becoming good companies, uh, in some respects. Because a lot of our customers today were also ideating on the same ideas, and then they did get somewhere with them. So we- I think one of the conclusions that I've pulled from it is, if you're very good, and, like, diligent, and can grind it out, you can probably get some idea somewhere, uh, to some successful point, so.
- ADAdvith
So then, yeah, through each of the iteration process, we learned something. Like, the first idea, we learned to not build something before we, like, have very clear validation. And I, I, I think for each of- e- each successive idea, we just refined our process for, like, how to validate an idea.
- RORobert
Yeah.
- ADAdvith
Yeah.
- 13:25 – 18:55
Forming the founding trio: rejection, misalignment, and the ‘don’t pick cofounders for the idea’ lesson
- HTHarj Taggar
Um, and then how did the three of you come together and decide to team up to find ideas?
- MAMarty
Yeah, so I had left Airbnb, worked with one guy for a year on a health tech idea. That didn't work out. Then spent six months, um, working with another guy, who is at- actually still at Airbnb, and then, you know, six months in, he's- it's clear he's not gonna leave Airbnb. So it's like, okay, back to the drawing board. I personally learned during that experience that co-founders matter a lot, and to be honest, I just didn't care about the idea that much anymore, personally, and I was just looking through my network of, like, who are people who I would just want to work with? And I'd lived with Advith for three years before, um, I considered him one of my best friends. And I was like, "Okay, like, we've never worked together, but this is someone I think it would be really fun to build a company with." And so I reached out to Advith, um, and then what was going on for you guys at that time? [chuckles]
- ADAdvith
Yeah, Ro- Robert and I were, were working together, and, um, it basically felt like since we're at this stage pre-idea, that it was actually counterproductive to have even more people trying to brainstorm together, and so it felt like not the right time, and it felt more like, oh, once anyone has something that might actually be a- an idea, let's all pile on, type thing. And so we were reluctant, for sure, at first, [chuckles] to, like, all work together.
- RORobert
We rejected him. [laughing]
- MAMarty
Well, well...
- HTHarj Taggar
[laughing]
- MAMarty
No, no, no, no.
- ADAdvith
Be truthful.
- MAMarty
It was, it was even worse than that to start because they were vague about it. [chuckles]
- ADAdvith
[chuckles]
- MAMarty
They were like, "Oh, yeah, sounds great," and then, like, crickets for, like, the next two weeks. And I'm like, "Like, okay, so, like, when are we doing this?" [chuckles] And they're like, "Oh, like, what do you mean?" [laughing]
- ADAdvith
[chuckles]
- MAMarty
And being super vague and, like, not direct communicators about the whole thing, [chuckles] which I feel like has been a trend for us continuing.
- HTHarj Taggar
[chuckles]
- MAMarty
But a- at least for me, I was like, okay, this is, like, uh, at, uh, the top of my list, um, you know, working with Advith, and then I started getting to know Robert through, um, our conversations. And they basically said no. I, I personally started just, like, continuing to work on ideas. Um, there were some that I tried to bring to them. Um, one actually that I feel like we spent way too much time in hindsight is the-
- ADAdvith
Met collection?
- MAMarty
... uh, no, the land idea.
- ADAdvith
The metaverse.
- RORobert
Oh, right. [chuckles]
- MAMarty
[chuckles] This was, like, when crypto, like-
- ADAdvith
[chuckles]
- MAMarty
... the metaverse stuff was, like, really big.
- ADAdvith
[chuckles]
- MAMarty
And so there was this, like, one investor in LA, and I was in LA at the time, who was, like, trying to... He was like, "If you guys go build a company where you, like, sell digital land in the metaverse, I will fund it." [chuckles] And so I was like... We, like, explored that idea for a bit, and we're like, "Okay, where could that go?" So yeah, just stupid ideas in hindsight.
- HTHarj Taggar
While we're on the topic, like, what did you all learn, would you say, during this process about co- like, finding co-founders, especially at that earlier stage when you're trying to figure out the idea to even work on? Like, what, what works and what doesn't?
- ADAdvith
We learned the same thing that he just said, which is, like, you shouldn't base your, uh, co-founders based on the idea. There were several iterations where we were, like, about to start this company in the logistics space. We actually feel like we validated the problem, and we, we felt like what was missing was, like, sales experience in the logistics space or some entrenched network. And so we were considering co-founding with these, like, two other guys who were just, like, deep logistics, like, sales veterans, and we just n- we actually just could never acquire, uh, end on an equity split that made sense, so we're actually quite close to, to-
- RORobert
I don't think we ever built trust enough in them, either.
- ADAdvith
Yeah, just, like, couldn't, and-
- RORobert
They were just too different of people.
- ADAdvith
Yeah.
- 18:55 – 24:12
Selecting a big market: working backwards from $10B outcomes and comparing ideas on a spreadsheet
- HTHarj Taggar
Um, one of the approaches I've had to sort of finding ideas is you start by, um, at least defining, like, the attributes of the type of idea you would want to work on. Like, um, Drew Houston used to give this talk at YC constantly about how he found the Dropbox idea by... He didn't know what the idea was, but he knew he wanted to d- build something that he would use himself, that was a deeply technical problem, that, um, monetized, like, from day one.
- MAMarty
Mm.
- HTHarj Taggar
Um, did you guys do anything like that when you were brainstorming? I'm just curious what, what your thoughts are on that.
- MAMarty
Yeah, I think we iterated, like, a lot, uh, in terms of the i- uh, like, what type of idea. So I, I guess, like, one component is, okay, we wanna build a really big company. Um, we definitely were-
- ADAdvith
What does that mean, though?
- MAMarty
Yeah, exactly. Exactly. So defining it more explicitly, these days, we'd tell you, like: "Okay, we work backwards from a billion dollars in revenue or a $10 billion company." Um, and so we actually looked at public, uh, companies, uh, public B2B SaaS companies at the time, who were worth over $10 billion.
- ADAdvith
Hey.
- MAMarty
And turns out, there aren't that many. Um, there were, at the time, 26 of them. Um, and so when you look at that list, you're like: "Oh, shoot, that's, like, not a lot. How do we, like, understand why they got to that size?" Um, and so-
- HTHarj Taggar
And what does that... How do you do that?
- MAMarty
Yeah, so I mean, this, this then goes back to, like, when we came back with our ideas, right? So when we f- when we started working together, uh, I had one idea I was working on. You guys had an idea. You were working on the logistics one. And this is actually, I think, such an important part of the story. The fact that we came together with different ideas and had to battle test them against each other was actually so helpful, because-
- HTHarj Taggar
Ah, ha, so you were, you were all sort of, like, researching your own ideas-
- MAMarty
Yeah
- HTHarj Taggar
... and digging into markets-
- MAMarty
Yeah
- HTHarj Taggar
... and then convening together-
- MAMarty
Exactly
- HTHarj Taggar
... to, like, debate them?
- MAMarty
'Cause then it was like, they were excited about their idea, I was excited about my idea, and then it's like: "Okay, which one should we work on?" And so then we literally- we spent, like, I think, like, probably half a day, one day, maybe, working on, like, the, the fintech idea I had. And, and we spent, I think, more time on the logistics idea, but then we literally made a Google sheet. I, like... I've looked back on this a couple times. It's really fun. Uh, with, uh, we started with two tabs. Tab one was this fintech idea, and we did a bottoms-up analysis of, "Hey, like, how big could this company be?" And so the conclusion was, like, if we, like, crush it, it's gonna be worth, like... or we will make t- high tens of millions in revenue. Then we did the logistics one. If we crush it, we're gonna make, like, hundreds of millions in revenue, and that's, by the way, the whole market. And then finally, we had this impression that, like, B2B SaaS, or, like, something that you could sell to any company, was kind of an end state for a lot of people who were pivoting. They try, like, all sorts of, like, v- verticals and then just, like, eventually pivot to some sort of horizontal SaaS category. Um, and so we took Rippling as, like, a model company and made, like, a third tab for them, and we're like: "Okay, like, how big can Rippling be? And what- how... If we did the same exercise for them, how big is their market?" And what we realized was actually their market is orders of magnitude bigger than, like, the things that we were looking at. Like, you just take, like, everyone who gets paid and multiply them times a subscription, and it's like, holy crap, that's like... It, it's clear why some companies can be so much bigger, is you will, you will inherently just be capped by the market you choose initially. And so then it, it was kind of like, I think there was a little bit of sadness for... It's always sad when you, like, invest into an idea and you're like-
- HTHarj Taggar
But-
- MAMarty
... oh, like you think it's gonna work.
- HTHarj Taggar
On that note, like, you were at Airbnb. Airbnb is, like, the classic example-
- MAMarty
Yeah
- HTHarj Taggar
... of the same timing with-
- MAMarty
Right. No, no, but, but the thing, the key thing here is, like, it, it's, it is an established big category, right? Travel is a huge category. And then, so and this is an important dimension. It's like, okay, there's the, the big category that exists already, um, and then there's some sort of emerging trend within it, which ideally we could latch onto and then grow with in an existing market. And so, like, uh, Airbnb is a great example of this. Um, cr- like, Bitcoin would be a great example of this, where, hey, you'd say crypto is small, but crypto is, like, part... It's like the emerging way of doing things, potentially, in financial systems-
- HTHarj Taggar
Like moving money around, yeah
- MAMarty
... money, like, and that's a massive market, right? And so you kinda latch onto the emerging trend within the really big market, and your bet is that that emerging trend is gonna grow really fast. And so for us, like, okay, we kinda looked at all these, like, bigger SaaS companies. Um, the, the biggest company on the list was Salesforce at the time, um, like, $300 billion market cap, and one thing that was really interesting... I can't say, like, it really played out exactly like this, but it's a nice story. Like, uh, Sal- [chuckles] it turns out Salesforce, like, their biggest revenue driver was actually Salesforce Service Cloud, which is their customer support platform. So everyone knows them for their CRM, but actually, their biggest revenue driver, um, I think as of the end of last year, um, $10 billion in revenue for, um, Service Cloud.
- HTHarj Taggar
Mm.
- MAMarty
And that, that's their, where they make the most money. Um, and so yeah, kinda looking at those companies, like, realizing we needed to be in something more horizontal, wanting to, uh, work on, like, kind of a, a big, a must-have pain point as well. That's, like, something we investigated a lot. Um, and then yeah, so big market, emerging trend, and then, yeah... Do you guys wanna talk about, like, how we explored from there?
- ADAdvith
Yeah.
- MAMarty
Yeah.
- 24:12 – 25:49
The ‘why now’ and the breakthrough insight: B2B support moving to shared Slack channels
- RORobert
I think it's also important to dig into, like, the why now, and that was very important to us, to try and figure out, like, why companies succeeded at any point in time. We like, like looking at all the anecdotes and all the different examples, like Salesforce having Cloud to beat out Siebel System, and that's why they ended up winning. For us, the why now ended up being realizing that companies talk to each other over Slack, but, like, we didn't even realize there were other why nows, uh, that helped us, helped propel us, and I think, like, now- nowadays with-
- MAMarty
... YC companies, AI is like the why now, but we didn't even know that would be a thing, uh, back when we were pivoting.
- HTHarj Taggar
So let's go, so, um-
- MAMarty
Yes
- HTHarj Taggar
... talk us through the story of actually arriving at the, um, Pylon idea, customer support, um, and maybe where does YC fit into that arc?
- ADAdvith
Yeah, I guess the way the idea came about is we were just talking to a bunch of people. One of our friends basically said they were facing this problem at their job. Essentially, this idea that they were talking to all their customers over Slack, which is a very new thing, and, like, that meant they couldn't, like, scalably reach out to all their customers, and they were losing track of all their messages. And for whatever reason, that felt like it resonated as like a, a kind of world view shift of, like, how people are operating, and it, it... Like, it was easy to draw the through line of, oh, a bunch of, like, the way you do your kind of customer processes would start breaking. And then we all gathered around that, hit up our whole network of, like, startup folks that we knew. It seemed like a bunch of people were actually at that very moment in market for a solution because they were like, this why now was a problem they were facing, and so that was very compelling. And the people who were not felt like they were on the right trend line. Like, you could see that in the future they would run into the problem, or they were bought into, like, this way of operating, which felt like it would lead to problems later on that we'd be able to solve.
- MAMarty
Yeah.
- ADAdvith
Yeah.
- 25:49 – 30:27
Discovery engine: LinkedIn outreach at scale and how they ran calls to surface real problems
- MAMarty
And, and tactically as well, like, one thing we did when we... W- we, we wanted to focus on one type of persona as well, and so we did try to look at, like, post-sales roles. I forget how we concluded on that exactly.
- ADAdvith
I think we thought, like, these type of roles were newer-
- MAMarty
Yeah
- ADAdvith
... and so it was, like, another why now depe- dimension. People like solutions engineers or customer success managers felt like newer, th- newer as a concept. Yeah.
- MAMarty
And, and so what we ended up doing, each of us, every day, we'd wake up, and we'd DM people on LinkedIn. At the time, it was, like, 40 max, uh, like, connection requests that you could send per person, and so every morning, that was, like, our first task. So 40 messages each, uh, personalized. We'd go on LinkedIn and, and just message people, and it was basically something along the lines of, "Hey, you know, we're founders who are pivoting around. Like, we- we're trying to learn more about, uh, solutions. Um, do you have 15 minutes to, to chat?"
- HTHarj Taggar
Do you have any opinion on... One of the common questions I get asked is, um, is it better, like, for exploring ideas, is it better to do that, "Hey, we're pivoting, we're trying to learn more about customer support," or is it better to say, "Hey, like, here's a solution. We'd love to sort of pitch you on it and, um-
- MAMarty
Yeah
- HTHarj Taggar
... and see how people respond?"
- MAMarty
I think it depends on where you're at in the, like, ideation process. At that point, we literally had no idea. So [chuckles] we were just like, "Okay, like, this category seems big. We're looking for the trend within it." And so, um, we actually didn't have an idea to pitch at the time. Maybe there were, like, a couple that we were, like, floating around. I think, like, when reaching out on LinkedIn, it's good to, like, put some little nugget in there to, like, pique their interest-
- ADAdvith
Mm
- MAMarty
... and then that in and of itself is signal whether they actually accept the friend request or not. Um, 'cause the response rate on LinkedIn, it was like 1 to 5% or something like that.
- ADAdvith
Yeah.
- MAMarty
It was, like, something abysmal, um-
- ADAdvith
Which is fine, though, 'cause imagine if you're messaging, let's say, like, 120 people, that's, uh, what's 5% of that? It's like-
- MAMarty
Yeah, still-
- ADAdvith
... o- over five people.
- MAMarty
Which is pretty good.
- ADAdvith
Yeah, that's fi-
- MAMarty
Yeah
- ADAdvith
... you're booking your calendar for the next week, right? And so, um, it- at that point, you can, let's say, like, talk to 30 people that you're trying to, uh, do discovery with the, the next week. And so, yeah, we were basically, as Advith said, looking for these, like, new roles and messaging them. Um, we knew we were in, like, a bigger, broader category, and then, yeah, kind of like listening for problems. And to give you a sense of what we did on these calls, we would just be like: Tell us about what you do. Tell us a- about what you care about, what tools you use. Like, what do you hate about your job? What does your boss care about? What do you measure? Um, kind of like getting, like, the full landscape. And so I think when, you know, Advith was on that call with, with, um, uh, our friend of like: Okay, you know, they have this Slack problem, or, like, there's some vague problem around it. Um, we also could, like, look back at all these conversations we had and, like, oh, actually, like, no one mentioned this as an explicit problem maybe, but, like, it's clear that other people, like, kind of mentioned this in the periphery of, of what they were talking about. And so it, it did start to come together, um, after that one person explicitly was saying, like, "Hey, this is a problem." More tactically, I think we... It's useful to spend, like, 80% of the call, like, trying to dig into, like, some idea you already had, but then leave 20% for, like, them to just say random stuff that could seed, like, the next thing that you think about or, like, that's like, you know, uh, creates the next set of questions you ask to the next person. And also related is, I think, at least for me, it felt like a lot of the leverage of quit- at this point, we had quit our jobs, uh, entirely, all of us, and so it felt like the main thing we got from that is actually the ability to iterate every day versus every week, which that cycle time affects the speed you get to your idea a lot, I think, or way more than I would have expected. I would have thought before I could balance everything. But I think the issue when we were working while still at a job is we would only change the questions that we asked, like, every week, for the most part, and we'd be, like, in the same headspace during the week. And so changing your questions every day just means, like, you're actually iterating every day.
- HTHarj Taggar
And you're doing so many calls at this point. You're, you're doing so much discovery work. Um, exactly what was it that people were saying was, like, the problem with Slack?
- ADAdvith
Specifically, what resonated for me at least, was that, oh, this is a new way... Like, this is changing how people talk to each other. Like, that feels pretty deep. Like, um, the- you never think of Slack as like a c- like a platform where you communicate between businesses. It's always historically been like, "Oh, I'm talking within my company." And so it felt like, oh, post-pandemic, everyone's moving online, so everyone's using these tools. Like, literally random businesses have to use some kind of tool to, like, talk both amongst themselves but also between each other, and so it felt like a big kind of shift in, like, how people operate. The specific pain point was like, "Oh, I have 100 customers, and I talk to them in th- over Slack. I have a s- shared Slack channel with them." Like, a simple use case of, "I have an outage. How do I let all these people know I have an outage if this is the way that I talk to them?" Like, "I have to literally go into every single channel and copy and paste a message." Like, you... The pain point is pretty visceral for even just that, even though it's, like, a pretty small slice of, like, what you would expect-... issues to arise, and if you assume this trend of people talking to their customers in this new way.
- 30:27 – 40:46
Early validation and ‘don’t do this idea’ warnings: passing the Mom Test and ignoring naysayers (wisely)
- RORobert
Yeah, it was pretty night and day in terms of passing the mom test. I think, like, all of the... I think each of us went to, like, one of our friends-
- HTHarj Taggar
For people who don't know, what, what is the mom test? And then tell us how it applied here.
- RORobert
Uh, yeah, the mom test is essentially, like, uh, if you ask your mom, like, "Hey, would you buy this thing?" Of course, she's going to say yes. So you have to ask questions around it to see, like, whether she would actually buy your thing. [chuckles] Um, but to back out into the, uh, our idea and how we thought of it, like, we each went to one of our friends' companies, and we were like: "Hey," like, we mom tested them, and it pretty much passed with flying colors around, like, yes, they would buy something there. They're actively looking for something. They tried to build it in-house but couldn't do it. Like, they went out and looked for other vendors, but they were all kind of bad. So, like, it, it was pretty clear that, like, we could find, like, a bunch of customers in these first set of people. I think also, like, we had originally... Like, it was kind of, like, right in front of our face, too, at, like, our past companies, where we had kinda seen this, too, from, like, the vendor perspective, where our companies had Slack channels with customers. Um, and we just never thought of it as, like, an idea until that moment. Yeah.
- MAMarty
Yeah. Also, I mean, um, to, to paint the picture here as well, so Advith, um, when he first spoke to, to, uh, this person on the phone and... Or spoke to this person who had this problem, um, he called Robert and I. Robert was, like, in a car ride from LA. I was living in a hacker house at the time. And so, you know, Advith's all excited. I think it's a Sunday night. You know, we hang up the phone. I, like, immediately... I was living in a hacker house called Mission Control, um, and I, like, went upstairs and spoke to another founder who was living there, Tejas, uh, founder of Hightouch. Um, and so I went up to him. I was like: "Hey, do you guys have this problem of shared Slack channel support?" He's like, "Marty, you're not gonna believe this, but we're actually looking for a solution right now."
- RORobert
Hmm.
- MAMarty
And actually, that they had worked on the same exact idea. So right after he told me, "Hey, like, you know, we're, we have this problem," and explained it, he was like: "Don't do this idea. We tried it. It didn't work."
- RORobert
Yeah.
- MAMarty
Um, and in fact, f- uh, fun fact about them, their name, Hightouch, is, uh, short for high-touch support, which was over Slack channel.
- RORobert
Yeah.
- MAMarty
So it's, um... So yeah, basically, it took i- 30 seconds to find, like, prospect number one after that phone call, which, uh, was pretty incredible.
- HTHarj Taggar
Since they had tried the idea, what's your advice to founders when they're... A common thing when they test out an idea is go speak to other people who have tried the idea in the past. Like, is that a good idea, or should they avoid those people?
- MAMarty
You, you should definitely... I think it- there are so many reasons why ideas cannot work out, and I feel like most of the time it's execution, or, like, they give up, or, like, there's co-founder drama. Like, I feel like most YC companies, you could t- say this better than we can, is, like, co-founders break up, right? And it had nothing to do with the idea. I think one great example of this is during our YC batch, it was the first batch after ChatGPT came out, like, the most common idea was, like, AI for customer support.
- RORobert
Yeah.
- MAMarty
And, like, almost all of those companies died. And, like, it's not like there aren't companies now that are doing really well as, like, AI support agents. If they had stayed with that, they could have been really big. Um, so I think there are all sorts of reasons. I counted at one point, I think we were... Are, like, the 13th company to try this exact idea.
- RORobert
Yeah.
- MAMarty
And in fact, a lot of them pivoted away and are now our customers for this idea-
- RORobert
Nice
- MAMarty
... uh, which is quite funny. So yeah, I think... Yeah, I, I think you should definitely, like, think from first principles if you think the idea is good. Yeah.
- ADAdvith
Yeah, I think you just have to have a reason that it'll work for you, that it didn't work for other people, and that is enough. It, it could be that, yeah, some team died, or that you know way more about the domain, or, like, your execution's gonna be better. Like-
- MAMarty
What was it for you guys with the-
- ADAdvith
The... I think it was actually timing was a big part of it, and I think our execution is, like, pretty solid as well. But, um, I think the timing was pretty big. The combination of, like, this trend of post-pandemic, a lot of these businesses are talking to each other, are forced to talk to each other in this, like, conversational way, plus conversational AI means you can, like, extract and make use of all of that data, felt like two great, like, kind of market signals. Yeah.
- RORobert
Yeah.
- MAMarty
Yeah, even when we were in YC, we were very worried about timing. Like, I feel like we... This was one of the existential crises during that three-month period that we kept coming back to of, like: Oh, like, is it too early? Like, is the market big enough? Um, and so we, we asked these questions. I remember we spoke to, um, during office hours with you and, uh, Dave Lieb, and he was basically saying something along the lines of, "Well, like, you know, th- this is a risk. That's right. Um, the market might not be big enough yet, but, you know, hey, that's, like, a risk I'd be willing to make." [chuckles] And I feel like that personally gave us-
- RORobert
[chuckles] It's great
- MAMarty
... a lot of confidence, um, that, hey, like, you know, oh, if a YC partner is saying, like, that is fair, then, like, yeah, that's probably valid for us. But that was, like, constantly something that we thought about, and even over the course of the next year, you know, our, our first product actually wasn't a big platform or anything like that. It was actually just a very niche integration to help customers who have shared Slack channels, um, ticket conversations from there into a ticketing system like a Zendesk or an Intercom. So we actually, to start, were just, like, this integration. We built it in, like, a week. We, you know, got it running with, with Hightouch, uh, during that same period. Um-
- HTHarj Taggar
Is it because that Hightouch said that, "Hey, this is the specific thing that we want?"
- RORobert
Yeah. Mm-hmm.
- MAMarty
Yeah, and so that... Yeah, and there were many problems around, like, this new conversation happening, like, over, over new channels. And it wasn't just Slack that we saw. We saw Slack. Microsoft Teams was starting to become really big, so at the time, like, M- Microsoft was investing a lot there. Um, you know, Discord was becoming a place where communities are. Um, WhatsApp, already around the world, that's where a lot of B2B communication happens. Telegram. And so, yeah, we definitely saw this trend continuing long term, and not just across one channel, but many. So it did feel like, okay, all these, like, business workflows and systems that rely on email traditionally for B2B, um, that same data model of email no longer works for, uh, kind of these conversations and new channels.
- HTHarj Taggar
You get the insight, you get the idea, you figure out the MVP v- for your first customer.
- MAMarty
Yeah.
- 40:46 – 48:37
YC entry story: using a two-week deadline to close a first customer and sharpen the pitch
- MAMarty
And then also, Robert, do you want to talk about how we got into YC? Because, um, you basically had a connection.
- ADAdvith
Yeah. [chuckles]
- MAMarty
It was kind of not the traditional path, actually.
- ADAdvith
It was a weird, yeah.
- MAMarty
I don't know if you remember this.
- ADAdvith
I remember this.
- MAMarty
Yeah.
- ADAdvith
What was the...
- MAMarty
Um, one of my friends was friends with one of the admissions people- [laughing]
- HTHarj Taggar
[laughing]
- MAMarty
... and said, "You guys are trying out this new thing for people who are pre-idea," and so they were like, "Do you want to talk to a partner?"
- HTHarj Taggar
Like office hours, like open office hours.
- MAMarty
Yeah.
- ADAdvith
Oh, okay. Yeah, we did a bunch of that.
- MAMarty
Yeah.
- HTHarj Taggar
So we get on the call, and it's not a partner. It's just him, and, like, the day before, we had found this idea, and we were like: "Hey, like, here are all the ideas that we've ever tried. What do you think of all of them?" Um, he ended up saying that the logistics idea was great, this idea was, like, okay, and all the other ideas that we had were really bad. Um, it actually lined up with what our gut kind of, uh, said, uh, about everything. But I think for us, we really wanted to go through YC with a, a validated idea and try to take it from zero to one versus, like, pivoting, because we already had built that skill of pivoting. And so that's where we were like, "Hey, like, can we spend, like, a week or two trying to close that first customer and build out the initial product, and then we'll come back and do the interview?" And so that's basically what happened.
- MAMarty
Yeah, so, so we came out of the call, and he's like, "Okay, do you want to, you know-... like, your ideas aren't that good, but we like you guys. Um, so do you wanna do an interview with YC, um, either tomorrow or two weeks from now? And so we said, "Okay, let's do two weeks from now," so that we could, you know, do that-
- HTHarj Taggar
Use as a deadline.
- MAMarty
Yeah. And so... And this was also Thanksgiving was in the middle of that, and so we were basically hustling to close, uh, Hightouch or, like, anyone else we could find. Um, and at the time, it's like we had, like, a Gmail address- like, we were, like, sending out calendar invites to, uh, Theresa, the, at the time, head of support engineering at Hightouch, and it's like: Hey, like, you know us, with our Gmail account, scheduling calls with her, um, try- no, no name of the company, no nothing, and she's actively looking for tools to solve a real business problem they have. So, you know, we show up on the call. We're both like, "Hey, like, you know, we can build this, but also, what do you need?" So [chuckles] you know, trying to, trying to play that game. And, um, yeah, she, she actually was evaluating other tools. Um, we beat out those other tools, which was awesome. Um, and so by the time we show up to the YC interview two weeks later, um, you know, we had... They were, like, just about to deploy it, um, our, our instance of Pylon. We unfortunately couldn't get them to do it over Thanksgiving, but they were, like, right there. I think they were ready to go. So we had a lot that we could say during the interview process. Um, we hop on. I think it was, like, 10:00 AM on, like, a Tuesday or Monday. Um, you, of course, are there, uh, with a few other YC partners, and for anyone who doesn't know what this is like, you basically... I don't know, is it still Zoom or no? It's Zoom, right?
- HTHarj Taggar
Uh, it's a mix.
- MAMarty
Okay.
- HTHarj Taggar
Most of the... We do more in-person ones now.
- MAMarty
So yeah, for us, it was, like, at, at the tail end of the pandemic. I think we were the, we were the first batch where things were gonna move to back in person, and so we hop on this 10-minute call in the morning. Um, they kind of like... You know, Harj is peppering us with questions of like, you know, "Why do you deserve to exist?" [laughing] Like, "What are you..."
- HTHarj Taggar
[laughing] No, you didn't ask.
- MAMarty
It's not part of the questions, but, but you know, it's... They're, they're basically digging into the business, and, like, one thing, um, that we actually had prepped a lot was there's this, like, online, um, Paul Graham simulator thing-
- HTHarj Taggar
[chuckles] Nice
- MAMarty
... um, where it's just, like, a bunch of common questions that you guys ask, and then it changes every 10 seconds. So the idea is, like, you have to answer in 10 question- 10 seconds and, like, rip through them, and so yeah, like, uh, we go through that. It goes by really, really fast. Um, and then later that evening, um, you give us a call, and you're like: "Hey, you know, we like you guys. Do you wanna do YC?" And I think at the time, we were- we still, like, didn't know for sure. At the time also, I think YC was going through, like, kind of a, a ch- shift as well, right? It was like pandemic, kinda changed everything around the world. And, um, and so yeah, we... Our thought process for doing YC, which I think is, is helpful for, for other people, is like, hey, YC fee- felt like much more than just investment dollars, essentially. And so, and, um, the... I think one thing, uh, that, that felt true was, like, every person we spoke to who was an alumni of YC was like, "You should do YC," like, "and YC's great." Everyone who hadn't done YC was like, "Oh, they take so much equity, like blah, blah, blah." And so, like, it's kinda telling that anyone who's actually done it is, like, saying you should do it, and then anyone who's not kind of probably, like, feels, like, insecure that they didn't do it, and then also, yeah, is just complaining about stuff that doesn't actually matter as much.
- ADAdvith
It's like a startup rite of passage, it, it, it almost feels like, so it's always cool to be part of that. And then I at least left the interview feeling like there's no way that, like, in 10 minutes, anyone can understand what we're doing. It just felt like a, almost, like, fake, [chuckles] like... Yeah, it's like there's no way in 10- But now I feel like interviewing so many people, I feel like I could tell in, like, 10 minutes. [chuckles]
- HTHarj Taggar
Yeah, your idea, you can actually tell pretty quickly.
- MAMarty
Yeah, yeah. And, and that's the other thing. I mean, this is same with, like, choosing co-founders is, like, the idea matters a lot less, and it's more of, like, how do they think?
- 48:37 – 50:47
From Slack integration to full platform: the product expansion path and PMF signals
- HTHarj Taggar
Let's talk a little bit about how the product's evolved, um, over time. You start off with like a, a Slack ticketing integration. Um, maybe actually, what is the product today? You know, what does, what does Pylon do in its entirety? And then maybe we can work backwards to how you arrived there-
- MAMarty
Mm.
- HTHarj Taggar
... and what were the signs of product market fit along the way.
- MAMarty
For the first year of the company, we were building this, like, uh, solution to solve the Slack channel support problem, where I have a ticketing system like Zendesk, I have customers who are in shared Slack channels, and I need the tickets to go into my ticketing system. So that's what we focused on for probably, uh, the, uh, most of actually our first year of the company. We got to probably around 400K, uh, in ARR with that. Um, and then we were thinking about where does the company evolve, where does it go? And one thing that was interesting was a lot of our customers were asking us for a lot more functionality that was not our responsibility. So we were very focused on just, like, this omni-channel piece, um, but they were asking for, like, adjacent features that were much more part of the core ticketing system. And so, um, we were kind of like observing that. We were observing startups starting to tell us, "Hey, I don't have Zendesk yet. I just want to use Pylon. Like, can I just use Pylon?" And we're like: No, no, you have to wait. You have to go buy this other thing in order to use us. And they're like, "No, if you just had, like, one more feature, like, you know, it's fine. I don't even need Zendesk." And so we started getting that feedback. Um, and then finally as well, Zendesk had been acquired by private equity just, uh, like, actually the day after we started the company, and over that course of the year, uh, people were starting to feel the effects of, um, that product not evolving. Um, AI's coming out. Like, the- like, it's not the modern product anymore. Like, in 2023, at the time, like, that is- that can't be the most modern tool, uh, to do customer support with. So all of those things started coming together of, hey, um, people don't actually like the incumbent that much. Um, people are asking us for more functionality, which we are realising is, um, all B2B specific, and actually maybe there's something around building, like, a support platform for B2B companies specifically. And then also, yeah, just AI is, is, is coming into the world. So all that came together. We decided, okay, um, we definitely aren't gonna build a really big company off of just this integration, and we need to expand. At the same time, our customers are pulling us in this direction of, "Hey, build, like, something that's way more B2B specific and actually the full platform," and there's an opening in the market to do that, so.
- 50:47 – 54:37
Becoming AI-native: skepticism, sequencing, and what works (and doesn’t) for B2B support
- HTHarj Taggar
Maybe on AI specifically, something I'd be curious to hear about, like, you guys, you guys sort of launched and started work on the product just before the ChatGPT really took, took off and, and really had its moment. Um, and you've adapted really, really well to becoming AI-native and really baking that into the product. Um, how have you done that? And then what's your advice for founders who are maybe in a similar spot to you? This, this might be like Series B or C companies who sort of built a solid SaaS product pre-AI and are now trying to figure out how they fit in in this sort of post-AI world.
- MAMarty
Yeah.
- ADAdvith
I think really early on, we were actually very sceptical of AI, and we were very resistant to adding any AI into the product. But I think, like, we just did the same product iteration around, like, listening to the customers, what their pain points are, and thinking about all the, all the possible solutions, and AI happens to be the best way to solve a lot of those problems.
- HTHarj Taggar
How do you separate out... One thing I've definitely seen people get pulled in is, like, customers are saying they want AI because, like, they're reading about it, and they're like, "We want AI stuff." Like, how do you separate out sort of that from, like, actual signal on something that will solve their problems?
- ADAdvith
People still do that today all the time for us, where, like, they want some, like, silver bullet solution of, like, just the canonical thing in customer support is just, like: "Answer all of my questions for me," or, "I'm, like, a two-person startup. I just, like, want something that just, like, answers [chuckles] all my customer questions for me." I feel like it's still not, like, a specific answer, but understanding, you know, where the technology is, where you are in the market. We serve B2B companies, so for example, in consumer, the types of questions, like, Uber Eats gets about refunding orders are way easier to answer. Like, a lot of the volume is in the same set of questions, and the margin for error is, like, much larger, where it's okay if you mess up five in every 100 questions. For, like, B2B, the segment of the market we were in, like, if a million-dollar customer asks you a question, you can never answer them wrong, and you never want it to feel impersonal. And so I, I feel like it's more about sequencing than, like, not doing something and waiting for either the technology or, like, market sentiment or, like, your own product vision to be in the right spot. And so at least today, maybe the high-level guiding philosophy could be there's certain workflows that make sense for AI to do entirely and should be agentic. Most workflows, I feel like, at least in our space, are, oh, it's a combination of people and AI that should be working together, and then kind of there's some things that just people should be doing, like building relationships, and then how do you, like, cater the product to that, like, world view? Yeah.
- MAMarty
Yeah, so I think it's, it's good that we were slower to... Well, we were basically more sceptical of whether AI is, like, the silver bullet for everything and much more approaching the problem of, like, yeah, there are many problems. Out of all the solutions we could build, like, which is the best, and AI just happens to be the right solution for a lot of them. Um, another thing to mention is our core data set with Pylon, if you think about customer support tickets, um, to start, is all conversational data. So, like, when you think about what, like, LLMs enable, it's like, okay, taking con- like, text and structuring it so that you can build workflows off of it or do thinking on it. Um, suddenly, that's, like, incredibly applicable to our core data that we, we house, and so that means all this unstructured data that used to be really hard to turn into structured workflows is now structural. And so that has been a huge unlock for us, where, um, like, it- there are just so many things that you ideally would have wanted to do off of, like, applying AI to, like, a conversation or, like, lots of data about a customer that are now possible. And so we definitely get to think from first principles on, like, okay, what are, what's, like, the, the grand vision or effectively, like, if you had a really smart, like, B2B ticketing system and CRM, like, what would you want to build? Um, and AI, like, turns out, can enable so many workflows within it-
- ADAdvith
And-
- MAMarty
... with this data, yeah.
- ADAdvith
For all the existing f- like, companies, I feel like they can do the same thing of just trying to introspect on, like, what is the core- what is, like, the core data source you have? Is it very, like, amenable to using AI? Like, what are the core workflows in your product? Which ones, like, have this low margin f- or, like, kinda can, can-... be fine with AI messing up sometimes, but still provide a lot of value, and like kind of bet on the technology improving and like sequence your product roadmap to like benefit from how that's gonna play out.
- 54:37 – 57:47
What’s next: expanding from ticketing to customer success/CRM, plus scaling the company intentionally
- HTHarj Taggar
Um, how about we sort of look towards the future, um, here? Tell us a bit about kinda what should we expect to see from, um, maybe Pylon, like the product, what's coming that, um, that you're really excited about? And then maybe just as a company, like how do you think about growing the company and the kinds of people you wanna attract?
- MAMarty
Yeah, I, I guess I can start on the, the product side and, and where we wanna go. Um, so we started with the core ticketing system, and, you know, this, this omni-channel wedge of, hey, Slack support team, support whatever. First off, that trend turns out is way bigger than we thought. Like, that is still one of the core reasons people come to us, so, um, that, that has continued to re- remain important now. But long term, uh, what's gonna be relevant is more and more customer-facing teams are going to build workflows, um, out of... in Pylon. They're going to live off of data in Pylon, and all these customer teams should be able to like work together off of conversational data. So, um, we, we started with ticketing. We're adjusting that well, building good workflows there. We're expanding into customer success and account management teams who want to handle customer data post-sale. Um, and then even adjacent teams as well. So for example, "Hey, um, I want to know all my feature requests." Um, and like the support team might wanna pass that data to the product team. So we can structure all of that data and give that to the product team now. Um, "Hey, the marketing team wants to know... um, we wanna do a blast to our customers who have asked for this feature in the past. Like, um, help me gather a list of those customers so that we can go send that data out." So, um, it's going to turn from like just core ticketing system into more almost like CRM. And so realistically, like very long term, this is gonna look less like a, like a AI Zendesk and more like a AI Salesforce. Yeah.
- HTHarj Taggar
How about the company?
- ADAdvith
We're going multi-region, so we're opening up a New York office. Probably we'll have a office somewhere in Europe as well later this year. Um, now we have like a V1 leadership team. We'll probably have like, um, several layers of folks within the company, and we were very [chuckles] resistant to a lot of these like normal company building things, like very, very resistant to like having managers or tracking work or like doing anything that most companies do.
- HTHarj Taggar
[chuckles]
- ADAdvith
But we're basically refiguring out from first principles or like running into these problems ourselves, why people do any of these things, and so-
- RORobert
And I think it, it was important that we didn't implement it all at once at the beginning, and we figured out why these things work. 'Cause I think a lot of people just like implement what they see at giant companies without really thinking like, does this actually push the company forward?
- MAMarty
Yeah, so, uh, it, the- [clears throat] yeah, so basically, I guess for the com- the, the business, um, we're still focused on getting to a billion dollars in revenue. That is still like the end goal and the, the win conditions for this game that we're playing. So yeah, um, if you like joined one of our, our meetings between the three of us, it's always like, "Okay, are we on track to do that? Um, what market should we be exploring? Um, you know, where are we gonna run out of, uh, space with the current like wedge product that we have?" And, and so yeah, um, still focused on a billion dollars in revenue. Product, uh, is really well-positioned to expand a lot, given our data set and w- where we live about around the customer and, and the, the roles. Um, so yeah, there's, uh, a lot of room to grow.
- ADAdvith
And maybe a main change is like basically being way, way more intentional about like literally everything that you might have assumed was happening just naturally before. So like being intentional about what we make product bets on, like where we hire people, what regions we expand to. Basically, all decisions, trying to become like way more intentional about them.
- 57:47 – 1:00:21
Founder growth reflections: people leverage, staying close to the work, and owning the ship’s direction
- HTHarj Taggar
Um, how about we finish up here? I'm just curious, does each of you have sort of like a, a reflection on, um, how... Like what's sort of one area, one specific thing that you've really grown in as founders, um, since you started the company?
- ADAdvith
You kind of hear this before, that like people are what make the company, but actually, like in our roles, like the biggest leverage is people. Like, just like h- you're essentially trying to use people to like do like bigger and bigger and bigger tasks, and just like, how do you do that, I guess? And just knowing that that's like the most important thing, I feel like I never internalized it until like we reached a certain scale where everything, like instead of turning to you doing it, then you have to turn to somebody el- You like operate in a different mode, where you're like actually trying to figure out how to get a broader set of people to do it, or who's the right... Aligning work to the right set of people actually cr- creates like a really big outcome if you do it in the right way. And so all these like people things I actually feel are what make the company, yeah.
- RORobert
Yeah, my answer is kind of similar to that [chuckles] actually, yeah. 'Cause I thought a lot of it would be like a lot of critical thinking on our end, of like, oh, we get together as a leadership team, and we like work on a problem. But like most of it is just as, as the tree gets bigger and bigger, you're further and further away from the IC.
- HTHarj Taggar
Yeah, you go from being like the person doing all the work to getting other people to do the work-
- RORobert
Right
- HTHarj Taggar
... do it well and consistently.
- RORobert
And it's still important to be able to dive all the way down and see what's happening at, on the IC level, but like, I think a lot of the work is actually just talking to people all the time, yeah.
- MAMarty
I think we are very in the weeds still with a lot of the projects, and I think that's good. Um, I think the, the longer we can stay, like, very involved, um, in, in like being evaluators and editors on work, I think, uh, will be very helpful for the company. And it is crazy to think, like when we think about company planning and, like, direction, I mean, it's all up to us. [chuckles] Like, there's like no one else that like-
- ADAdvith
Yeah, I guess-
- MAMarty
... has the context
- ADAdvith
That's another realization.
- MAMarty
No one else is, like, paying attention in the same way. Like, you might think, like, oh, investors over time are gonna, like, be able to, like, guide you of whatever. No, it's, it's just you. Like, you [chuckles] like, like the three of us right here are going to determine the futures of all 79 people, and like-
- ADAdvith
Yeah
- MAMarty
... everyone who joins. And like, if we take our eye off the ball or stop working as hard, like, yeah, it's, it's... We cannot. Um, and so I think, like, yeah, there's, there's the founder mode piece to make sure that we're involved, so we can make sure the right, like, uh, decisions are being made internally, and then also kind of the, the higher level, like, we have to steer the ship because no one else is gonna do it. And yeah, yeah, that determines the, the long-term fate of the company.
- HTHarj Taggar
Nice. That's the exciting part of being a founder.
- MAMarty
Yeah.
- ADAdvith
Yeah.
- HTHarj Taggar
Cool. All right, I think that's all we have time for. So thank you so much for joining us, guys. It was really interesting to hear the story.
- ADAdvith
Thanks, Harj.
- RORobert
Cool. [upbeat music]
Episode duration: 1:00:21
Install uListen for AI-powered chat & search across the full episode — Get Full Transcript
Transcript of episode XcI5IMLEAaE
