Vamsi Saladi, Contour: CTO @ Contour (YC S23)

Welcome to another Founder Wisdom Pod YC edition. We have MC Salari with us today. He's a CTO at Contour, also co-founder.

Contour is a YC uh summer 2023 companies. Vampy's in uh New York and yeah, it's a pretty cool company. It's um a voice agent companies and it's for outsource uh call centers and I had one call centers uh with 200 employees three to four years ago.

So it's going to be pretty fun one. A quick one about um BAMC has been a data scientist for Rocket Ship VC which is quite an interesting one as well. We'll talk about that one.

Software dev at Hibshot. Um course assistant at Stanford. You've also worked at AWS.

So big u experience there. I would sum it up as data science and also AI. So that's going to be a fun podcast today.

I do a good uh intro. Can you dad do a better job at that? No, no, you did a great intro.

Um, yeah, I' I've been a little bit all over throughout my life, but you know, I did Rocket Ship, which was more like the VC world, Amazon, Big Tac, um, Hype Shot was a startup. So, I think I have a good flavor experience coming into this, right? And why launch your own company?

Um, I think because of the experience where I've kind of tried all the different things before. Uh, I enjoyed my time working at a startup way more than anything else I ever did. Um, the VC world was cool, but I think startups are much more fun.

And which of the startups you worked that proved to be a very good investment of your time and what did you learn there? Yeah, so I think there were two startups where I had a lot of learning experience and also just a lot of like growth and mindset. So the first one was Polar.

Um that was around 2019 I want to say and I was like an ML um intern there. Um they were kind of like a computer vision startup and I worked on like a really cool project or what was a cool project in 2019 I guess but it was the idea that like if you have group photos of people and some people are looking one way some people have their eyes closed whatever it was building like a vision model where you could output an image where everyone is looking at the camera. Um, so that was really fun.

I learned a lot and learned how to learn in like a small group environment, you know, like fast iteration, not much red tape. You can kind of just do what you need to do. Um, and then the second one was Hypershot, which was much more recent, and it was actually a company started by one of my like good friends from high school.

And he was just like, this is during COVID. And he was like, "Hey, like we both enjoy sports and like I'm building something for, you know, sports streaming, like fantasy sports. Um, do you want to hop on and like help me out?" And I was like, "Yeah, of course." Um, I'm a huge sports person.

And so I was like, "Yeah, let's do it." Learned a lot over COVID. uh lived briefly in New York while working with them. And I think both of those together kind of told me, okay, this is where I want to be.

I don't want to be in like Amazon or I don't want to be in BC. But still like very scary. There's so many unknowns and it's a totally different world, right?

You said you're a sports person. Do you enjoy better watching sports or playing sports? What What do you play?

Um unfortunately the startup life isn't great for playing sports. But I do enjoy playing. Um, I do enjoy playing.

I played a little bit during like high school and then didn't really play during college, but I was like a martial artist. Uh, I played soccer and then I was also a gymnast for a little bit. And I still enjoy doing those, but right now I mostly watch.

Right. And the describe the the sport that is running at a YC company. Is it like 16 hour days?

What a day in your life look like? Um it's it's rough. I especially the past like I think week or so it's been extremely busy because we're like starting to onboard some people and it's just been a hectic you know process but um honestly it helps that I actually live with my co-founder and so we can you know very tightly align and make sure we do everything together.

Um but yeah we will usually wake up together around like 8 or 8:30 in the morning and from that point onwards it's just you know work. Uh we have like a little bit of time set apart for lunch and dinner, but we usually wrap up around, you know, 12 or 12:30 and then So, how do your neighbors describe you? Okay.

Two quiet guys that hustle all the time and play music out sometimes and shout outs. No, no, no. I think most of our neighbors think we're pretty quiet.

Um because we don't we're not really like we're pretty like locked in. I think we're we're pretty focused. um we don't have like a ton of like noise around us when we work and I think we prefer it that way.

Um there have been times in the past where we've worked in like co-working spaces. I think those neighbors might describe us differently and you're building and he's raising selling like how does that work? Yeah.

Um we have a pretty good separation but it helps that both of us are technical. So actually for like a good bit of time we both built and even like now we both kind of build. I think yeah if it comes down to it technically I do more of the building but and he does more of the sales more of the marketing more of the fundraising but both of us can code so that's a helpful thing to have vibe coding a lot nowadays um not really I I I feel like I'm a I'm not a huge vibe coding person why is that this is going to make me sound really old but I think there's been Yeah, there's something about there's something about just like knowing what you're coding being prototyping like planning it out and you know just like writing code yourself.

Obviously like we still use cursor, we still use cloud, we still use uh GPT but it's more as an auxiliary tool and not as like the primary driver of productivity. And maybe you're just not at this stage of vibe coding which is more prototype, right? You've been building this for two years already.

So you have a solid code base. Uh how many times have you pivoted since you started in 23 or has it has the concept remained the same? Uh no.

So there's been one big pivot. Um what we went through YC with and what we raised our initial preede with was actually one idea. And so that was more in like the QA dev tool space.

And then this kind of voice uh support platform was born around let's say five or six months ago. And it seems to be to me that there's a couple of voice companies doing that right now, right? Like are you keeping track of them?

Are you like coating at frenemmy sort of? No. Um, I think one of the things I like about kind of other YC founders is for the most part we're pretty transparent.

Um, because we, you know, we're part of the same community and so we we can see what everyone else is doing. Um, I don't think the space is a winner take all and so we're happy to help each other and kind of work together. And what are you building now and what makes you different?

Do you feel you have a product market fit at the moment? Yeah. Um so we're building an AI customer support platform specifically for financial institutions.

So think banks, credit unions, BTOC fintex. Um it's an omni channel support platform where our flagship product is the voice. And so we you know emphasize that a lot but we do kind of do the other channels as well.

And I would probably say that we're like, it's still a little premiumf. Um, you know, we have, you know, people interested using pilots, all that stuff, but I don't think we're at PMF yet. So, is it voice and chat all in one?

Like, uh, then I see your ideal customer person would be a bank, think tax and credit unions. Can you give me a use case? Let's say I'm a bank like how would I use you precisely?

Yeah, sure. So, for example, if say you have like a support staff um because you're like a pretty big bank um you can use contour both as a co-pilot but also in its autonomous form. So maybe for say you have like a phone line or you have a call center um and you're paying them for like say 9 to5 Monday to Friday.

Um you can enlist contour to be your support line over the weekends or during the off hours. Um, if you already have a support team, then you can use the Contour Copilot where it can help, you know, help help them be more efficient in how they answer tickets and call tickets, but it doesn't like technically replace them. Um, and then if you're a small company, if you're like a series A startup maybe and you're like a B toC fintech, um, and you don't have a support team, then you can just use Contour for everything.

It can be your chatbot. It can be your, you know, voice line. It can be your email support.

Uh, and just kind of handle everything for you. And who are the big players you'd be replacing? Or are these banks not even using uh SAS?

They're using their own custom platforms. Yeah. So, there's two categories.

Um some some people do have their own you know custom platform but for the most part people use like companies like intercom um which are also like support platforms kind of more geared towards financial institutions um or they use Yeah. Yeah. Zenesk is more like the ticketing system but it has some AI features but it's not really not really that good.

Um but yeah, it would be kind of in that vein and we don't necessarily want to replace the ticketing system. So we can work with Zenesk for example. Um we want to replace the actual support.

So that's usually either their or internal support team or they have like an external call center or just outsourced you know support staff right and uh in the two years you've been building how many clients so far and how do you determine which new feature to launch as the CTO? Um so the first part uh this is more been only going on for around five or six months or this version of the product I mean. Um so the old one we had like we had customers we like raised a pre that was like a whole journey on its own but this is a new one.

This is a new journey. Um this one we have a few pilots right now. Um no one's in production yet but we feel pretty good about kind of where we are with the pilots.

Um, and then in terms of what to build next, I think that's always the tough question, but I think thankfully we're in a space where this is something a lot of banks want and a lot of fintexs want. Um, they don't know exactly how they want it. They just want the problem to go away.

And so that gives you a little bit of leeway as like the engineer where you can be like, okay, this is the way to solve the problem. um or this is the way to solve the problem where you don't have to be like oh I have to build exactly like this product or this feature or you know this functionality and are you back in that ramen days nowadays does that feel sort of depressing or do you feel do you still have fuel in the tank I I think have a lot of fuel in the tank right now u we have this like new shift has given us a lot of momentum I think there's a big difference when you're working in like the dev tool space and when you're working in like a like the fintech world where you're selling more to like slightly bigger companies like it's more driven you're solving like big problems for them. Um as soon as you get like momentum it just feeds on itself and I think we've gotten a lot of momentum and so we're just you know riding the wave and seeing what we can do.

And how big are these problems? What um paycheck would the bank be willing to send your way on a monthly basis? Would it be on a per usage basis on a Yeah.

Yeah. So, it's usually an annual contract, but it is like usage based in the sense that we will estimate based on the pilot period what your volume is going to be over the course of the year and then charge based off of that. Now, if you're over under, you know, we can reimburse or charge overage and that's fine.

Um, but for banks, I think it obviously depends on the size of the bank, but maybe like a smaller regional bank or community bank, somewhere between like 50 to 100K is like a expected, you know, contract. Um, yeah, sorry. Go ahead.

Do you need uh do do you do uh demo calls yourself or is that your co-founder? And you actually need to meet with these banks, right, which most of the time are old people. Do you need to meet with them physically in New York?

Um, it's true. We've worked with a lot of older folks, but um, no, we both do demo calls. Um, I think we're both comfortable selling and we're both comfortable coding and so I think we both do whatever needs to happen.

Um, in terms of meeting the banks, it sometimes it helps. Um, but it's usually before the sales process. Um, I think doing discovery or getting the initial lead in person makes a huge difference in this industry.

Um, because a lot of it is just about like are you willing to trust like me as a founder slash our company to even get started because once you get started with them, this is a huge need. Uh, this is a problem they want a solution for. And so they're not they're not going to not talk to you because it's not like important to them.

They're usually not going to talk to you because they don't trust you or think you're too small or like have some like other concerns. Um and so meeting them in person helps a lot with that. Yeah.

Very low churn, I would guess. Does your solution replace employees or amplifies them? Because I guess that would be at the friction point with them.

They're like anti- AI sort of folks. Yeah. Um that's one of the biggest lessons we learned I think throughout our like startup journey is you have to find the right person to sell to.

Um, so for the old product, we sold a product to the person that it was going to replace and naturally they were like, "Oh, we don't want that." Like I don't reality. Exactly. Yeah.

Um, and so we learned this time and so we usually target kind of higher up in the chain and for the most part the actual product does save like it's like a cost cutting tool. You don't have to hire more people. You can replace up staff.

you don't, you know, you don't need to pay for a call center. Uh, but it also is kind of like amplifying for the current staff because we have a co-pilot as well. And so, because we offer both, it's like a double whammy where you can both be more efficient with the people you have and you can, you know, maybe not hire more people.

With your first pivot, uh, how did you manage investors expectations? Because I mean, they know that YC companies pivot a lot, right? So they know that their money might be invested into a vehicle that might change course with time.

Were they just down with that? How did you manage that pivot? Yeah, that's a great question.

So I think our investors were incredibly understanding and yeah, as you said, I think it's expected from YC companies to, you know, kind of pivot around until they find the sweet spot. And so with the first pivot, we ended up doing that product for almost like a year and a half, right? And so we put in a lot of time, you know, we had customers, we had usage, like it felt like the the real thing.

And so when we decided to pivot, it was a tough like tough decision. U because you're leaving like users on the table. You're leaving like money on the table.

Um but I think our investors realized the same things that we did, which is we're selling in an industry where every single sale is like an uphill battle. The need like the burning need is just not there yet. And so everything felt like grimy and forced and it was like so much effort to like get people to use it.

And going from that to an industry and a problem where everyone wants the solution and as long as you're like performant and like compliant and relatively cheap like anyone will pay for you. It's like a huge difference in mindset and perspective. And so our did you know that like was it that YC mentors telling you that and you seeing YC peers like see making two three 10x more than you do?

Um, I think it was partly it was partly us knowing that voice is like it felt like the next big thing, you know? It was like this was six, seven months ago where before all the new like Sesame or like all the new models and 11 Labs and all that stuff like when Voice was still pretty young, we were like, "Okay, this feels like a space that's about to blow up and we're still pretty early into that journey. So, there's obviously a lot of opportunity." And then when it came to the vertical like because you could do general voice, right?

You could be like a vap or like a bland. Um but in terms of like where we wanted to focus or what we wanted to solve. I think we just had some friends in the like fintech world and they were like look I sell to fintex and this is a problem for them.

Like they've told us this. They solve other stuff but like if you guys want to tackle this problem like go for it. And so we were like okay we see the opportunity.

We like the voice like world and we like the technology and we're both like we have experience in it and so we're like let's give it a shot. You have good friends man. They give you top.

Um Sesame is crazy good right? Like I can't even compare it to Chad GBT's voice at AI. It's so crazy good.

Um what are you seeing in in that space as well? What's the other company sent something? Uh what's uh what's exciting to me is that I could have agents call for me.

Uh I could even do this podcast on a phone call, right? I could call people straight up and be like, "Hey, I'm doing a pod. Like, do you want to record it right now?" That that's exciting to me.

What else are you seeing in that uh space? What kind of features? What kind of companies are leading the way?

Um Sesame is a really good one for sure. Um, I know they got into some like trouble because like they like fake open sourced it, which is kind of annoying. Uh, but I still think, you know, they have a lot of potential there.

In terms of, you know, other companies, I think there's like the whole like horizontal approach, right? Like if you're going to be like a bland or retail or a VP, um, these are all companies that are trying to be like the voice provider and they're all successful kind of in their own way. Um, and then there's more like verticalized companies that are doing voice but like in very specific use cases.

Um, like there's a VOCA is one of our friends and they do like voice agents for HVAC units like one extremely specific use case, right? Like it's you wouldn't think it's like a massive industry but like it is. Um, and so they're able to focus on that one vertical and also be successful.

Yeah, dude. The age back in the US it's humongous. like these are good old schools like salesmen knocking on your door and so forth it's so big um yeah I guess that's a YC model as well right like crack one specific niche because my next question is like why don't you start your own vap or sesame I guess it's like folks that are at different stage of their careers know a bunch of investors as well they they have like huge experience at open AI for example and uh yeah why don't you build your own is it like make a bunch of money make cash flow flow satisfy investors and eventually uh start a new segments.

Is that how it works? Kind of. I think one of the dangers of trying to solve like the whole problem is often times you end up solving nothing.

And so it's like a weird balance to strike there. And so we wanted to be very kind of focused and determined on what exactly we solved. And so we chose specifically consumerf facing like financial institutions.

So, banks, credit unions, uh B2C, FinTax, and we chose specifically that we wanted to help with customer communications and support. And so, we do like compliance workflows for example. Uh we have like a bunch of like policies and securities in place so that they can be sure that like you know data is good and like communications are going to be good.

There's guard rails on what the agent can do. Um they can trust it to handle like a lot of volume. And there's so many things that are specific to this industry.

So for example, if we have a chat agent, it's mostly useless unless it actually solves the issue for you or solves the issue for the end client rather. And so it needs to be able to take actions. So that's like integrations with digital payments like banking course um you know like refund a payment or um like enable bill pay like instead of calling Yeah, exactly.

like instead of calling in to like Chase or like Capital One, um imagine that you can just like call into like the boys agent or even like use the chat agent and just be like, "Oh, I want to set up like bill pay on the second um for like X product." And you can just like type that in and like the support agent will just go do that and you don't have to like there's no human involved like you don't have to think about it. It's just done. And so we want to solve that problem.

It's a humongous problem. Like your friend said, I don't even think you needed your friend, right? You're just a customer of Chase yourself for example, and it's a terrible experience, right?

Like you need a codes there, you need to call, you need to wait. Uh so it's just a a very inconsistent experience uh throughout and most banks um know they're probably losing customers, right? But even if I run away from my bank as a customer, I'll go to another bank that haven't tried customer service, dues.

It's like it's more of the same. It's it's throughout the industry. That's a huge opportunity for you.

Yeah. Exactly. And so we're very excited about like this space and this problem because it's an industry where there's a ton of people like each bank so many customers that it manages.

Each like credit union has so many customers it manages. And so the opportunity there is quite massive. And there's so much of like there's like an entry barrier because you need to get and establish a ton of trust in the space because it's like it's money like it's like the most sensitive thing you can like handle and right do you need to be compliant on your end?

Do you need like certifications and that silly question right? No it's Yeah. There's a ton of like restrictions and rules that go into place about like what we can do when we can do them what we're allowed access to and not allowed access to.

I think the good thing about working in the space is once you've established like the compliant nature of your product and once you've established trust with your customers, it goes a long way and it's really hard to like switch off of you then. And who's in charge of the compliance or customer trust part of that? And was that more frustrating for you than to build code?

Honestly, a little bit. Um, so the actual sock to compliance my co-founder is handling and so he, you know, he probably has more, you know, issues and problems with it. Psychotherapist bills.

Yeah. Um, it's it's a it's a I know a good psychotherapist in New York, by the way. One of my clients.

Shout out to Nick Samstag for anyone dealing with compliance. Um, okay. Maybe I'll need his contact just in case.

He's in the financial district, too. I'll give you his contact after that. But yeah, one last question because I know we're we're on time already.

Um, have you cracked a 50 to 100K deal? What's next for you this year? Yeah.

Um, I think that is um we are in pilots for those deals with like a a few different people. And I think if we can, you know, close everything, it's going to be such a massive stamp for us. Um, obviously our next like real goal because we feel comfortable with this is kind of cracking like the 20 million and up kind of bank, like the much larger bank.

Um, maybe not the level of like Chase, but like, you know, in between Chase and like a local bank. Yeah. Uh, so that's that's kind of the goal and we want to, you know, slowly work our way up the the size chain.

Where can uh people find out more about GMC? Yeah. Um you know follow us uh follow our company on Twitter and LinkedIn and then also follow me on Twitter and LinkedIn.

Uh I'll have the you know the links up when I post this but u you know keep track of you know our journey. Uh we're planning on putting out a lot more content both with you know people like you who host these like great podcasts for founders to talk um and then also just like kind of our own content because it's easy to demo you know our voice agents. It's like fun.

It's flashy and so we're going to try to do more you know funny content and marketing content as well. There you go. Yeah.

Shameless plug on my founder humor newsletter that's quite fun for that. I have an algorithm, a GPT for you to to use. And yeah, follow Vansy on Twitter.