Emergent: The AI App Builder for Everyone

I think AI is a big reset, right? I feel

that like uh it's going to be the next 20 years is going to be about AI. It's

just we're just getting started. I tell

people that this is Bitcoin and dollar one right now. And so you should just be bold and and you know uh find a problem that you really like and just go go all in.

>> Today I'm joined by Mukund and Madav Ja, the founders of Emergent. Emergent

recently announced a crazy 15 million error just 3 months after launch. Before

diving into that story, let's get to learn about Emergent. Can you tell us more about what Emergent is doing?

>> Great to be here. Uh Emergent is an AI app builder. Uh we allow anybody uh

app builder. Uh we allow anybody uh without any programming knowledge to come on our app and build a mobile app, web apps, websites with zero coding knowledge. You come in, you prompt and

knowledge. You come in, you prompt and you get a full production ready uh launchable apps. would just describe

launchable apps. would just describe what they want to do.

>> Yeah. They just come in, describe, >> you create the app.

>> Yes. Yes.

>> Awesome. And can you tell us a little bit more of your scale? So we know 15 in 3 months. So we launched 3 months back

3 months. So we launched 3 months back and have been growing pretty rapidly.

Probably one of the fastest growing uh AI startups right now in the world. Um

and we uh in 3 months we have like roughly 1.7 million users on the platform. More than 2 and a half million

platform. More than 2 and a half million apps have been built uh on the platform already. And uh

already. And uh >> that's insane. Actually live like in prod used every day. Yeah, I mean I think we we don't have a good track because a lot of people also deploy outside of our platform. We allow them to deploy on our platform, but roughly

20 30% people are able to deploy their apps on the platform today.

>> That's awesome. Tell me more about like how did you come up with that idea? Like

you were two brothers, right?

>> Yeah. Yeah. So we are twin brothers. We

actually uh got started. We came to US to do our PhDs. Uh and then I dropped out of the PhD program. Joined

>> and you went back to India.

>> Yeah. I joined Google then went back to India. Started a company called Danzo

India. Started a company called Danzo which was one of the pioneers in quick commerce in the country. And that was really Jack. How many employees did you

really Jack. How many employees did you have?

>> We had like big company like 3,000 employees but like 80,000 delivery people on the ground. Um yeah and uh so September 23 is when I left uh Danzo

came back to us. Uh spent some time with Maddie and we were thinking what to do and and that's when sort of you know we sort of emerged and Maddie >> I finished my PhD in computer science and then moved to the west coast and uh

got the breeze of the west coast for for doing startups. So joined Zeness as an

doing startups. So joined Zeness as an early employee and saw the rise and fall of Zen to some extent.

>> Yeah. And then uh went and joined Amazon uh was part of the their deep learning platform team built their SageMaker platform and yeah like uh last year me and Moken connected and so we we we've been wanting to do a startup for a long

time. It just happens.

time. It just happens.

>> How is it to build together like as brothers? Like is it a good thing or a

brothers? Like is it a good thing or a lot of fun?

>> How much tension?

>> We are very tight. We are very tight as twins. uh we are very close and uh yeah

twins. uh we are very close and uh yeah >> I mean we we have spent hours just chatting about ideas like every time we get free we'll call each other and say hey I have a new idea and uh >> it's more like a dream come true then

>> totally totally it's fine and you know like we've been meaning to do something together for a long time and finally we got a chance and uh I think the best part is that you know like we in most meetings it's just either one of

us need to be there and because we we sort of you know know each other really well so there's very high trust >> and so let's go more on the idea itself because you did y in summer 24, right? A

little over a year ago.

>> Uh but you didn't start with this idea, right?

>> Tell us more about like the the journey here, >> right? So when we came to IC, our first

>> right? So when we came to IC, our first idea was uh that we wanted to build testing agents which will actually go and test your web app, mobile app. We

thought browsing agents were getting really good and can people just in natural language prompt what they want to test and you know our agents would go in and and test that >> and from the beginning we had this very uh strong belief that you know the AI

exponentially is going to continue and you'll have a lot more autonomy uh in agents coming on so even though at that point like co-pilots were like the popular trend a lot of people were building co-pilots we were very focused on building agents which can do long

horizon task um and when we started like we there were two three things that happened one like we built out the entire tech and we realized that um the general coding agent is much bigger problem uh to solve and we were a lot

more excited by we felt that was a more ambitious idea that that felt like >> I remember like felt like that kind of like when you were were speaking back then.

>> Yeah, I think we went back and forth on multiple things and and and frankly we were still trying to find our way you know NYC when we were there to to to see what would work for us >> and we felt that general coding is much more ambitious idea with sort of you know something that people will be

excited about the next 20 years building. uh but then you you focused on

building. uh but then you you focused on more the enterprise market for a while.

>> Yeah. Initially our focus was enterprise market. So we built uh one of the

market. So we built uh one of the world's best coding agent at that point.

Uh this is this benchmark called Swebench and I remember texting you that hey we we we we are on number one on that benchmark now. So we attempted that benchmark. It took us like about a month

benchmark. It took us like about a month month and a half to uh you know crack that and become number one in the world on that benchmark. And and during that phase like I think we did a lot of like

uh research work between me and Maddie to sort of really figure out what would work in agent. How do you build long >> I mean even with an API access it's there's a lot of things that goes behind making those API access work for you in

in an agentic manner like so you want the agent to have very auton autonomous work and so one of the insights early insights we had was that u you want to give agent feedback and a very tight feedback loop and so that was one of the

genesis of emergent where we built all our infra inhouse right so uh the entire uh infrastructure be it like database back end everything like is built inhouse and uh that ensures we a very

tight feedback loop with the agent and uh we did bunch of other things around agents like being making making it multi multi- agent tech and and things like that.

>> So you were building all of that first for QA then for enterprise >> and by the time you went to consumers like kind of more of a consumerf facing.

>> Yeah. So we went to we went we went to an enterprise and and found the feedback cycle to be really slow and I think in the in heart like we are very consumer as a as a as a you know like

personalities having been a consumer company before. Yeah, Denzo was pure.

company before. Yeah, Denzo was pure.

was pure consumer company and uh so it felt like you know like we were getting too tied up in in you know like enterprise cycle and um and and we were internally building a lot of apps using emergent already right and so we thought

okay you know like why why not just go to >> they were like what you did for one year before launch is kind of what made you successful once you launch kind of like you were more ready than ever >> very much I think I think the foundation

that we were able to set in the first like 9 months of you know like just pure research around agents uh and uh really got us, you know, like u really into the problem and and helped us understand

like how to build really good high quality agents and and that's what's driving a lot of our growth today like when people come to the platform and they can see the difference between other platforms that we're able to build better apps.

>> That's cool. So let's speak about the product then uh like so after someone writes a prompt what does happen in the like from the time they click enter right >> what does the product do?

>> Yeah. Yeah. So behind the scenes as soon as you enter the prompt right like we spin up a devbox kind of environment which is a virtual machine uh in the cloud for for the agent. So the

hypothesis is that agent should basically get a same set of uh environment as as a developer would get right and uh in that environment agent is able to install libraries you know uh

write code and then like anytime it runs some some some like let's say write some files we automatically run some lint passes around it and give feedback to it. Um and and so I mean actually I

it. Um and and so I mean actually I skipped the step where like we we do uh first understand what the what the human wants right cuz people come from all walks of life and so we first have to like do so one of the unique things we

do is we first interact with the >> so you actually have like some conversation before actually starting >> before starting and also based on the conversation we realized where to route the right uh request right it may be

that this person is mostly interested in a pure front end side of things or this person is more interested in uh you know like via mobile apps or or any of those things.

>> So, so today you are able to do both kind of like a web app, a mobile app, like different kind of >> Yeah, we are the only coding platform that supports web, mobile back end, everything integrated in one place today.

>> And so you're going to root towards the right agent. It's different agents if

right agent. It's different agents if you have the tasks.

>> Yes.

>> I mean there is a lot of vibe coding tools out there. How do you do you differentiate from them? Like what do you do differently?

>> Yeah. So I think the way we differentiate is that most uh platforms today like help you get a good prototype, right? So they'll help you

prototype, right? So they'll help you get uh front-end code really well and and and our sort of goal has always been to take a user from an idea to a launched app that they can get usage on,

get monetization done on that. And

because we have we have tightly coupled the infrastructure and and state-of-the-art coding agent and the product like we have tightly built these things together, the overall experience of building the overall quality of the app building is is much higher on the platform because we have all the

feedback loops tied up. We don't depend on third party for your backend databases and that allows us to build really high quality apps on the platform >> and and do you do anything differently like to get to that level of production readiness?

>> Yeah. So so I mean at the heart of it is is a multi- aent architecture that we have built like where different specialized agent will come in and >> including testing >> including test. So we we do autonomous testing for you. We have a design agent

that will come and design your app for you. We have a security agent when

you. We have a security agent when before deploying it'll do security checks for you and and then there's a deploying agent which will convert all of your build steps into you know infrastructure as a code and deploy them

into uh into our cluster and and that really uh gives like a very like unified experience to to people to sort of you know build complex apps today.

>> What would be the limits today kind of like >> yeah so today like I mean typical I mean line of code is a bad metric but uh you know that's one one metric of complexity. So today like we typically

complexity. So today like we typically see the apps build roughly 35,000 40,000 2,000 line of code. Most of the other platforms sort of cap out at like 10,000 line of code.

>> So you think you are going beyond those platform but you still have a cap.

>> Yeah big yeah 100,000 to 100,000 line of code is where you know platforms but we are expecting as we add more platform uh improvements um as the models improve like we expect this the ceiling to keep

going up. So what initially it was just

going up. So what initially it was just kind of small apps toys already today it's kind of like full-fledged app more you could go even beyond that.

>> True. True.

>> Do you have like maybe any favorite apps you can tell us about?

>> Yeah. So I mean lots of apps like I think when we talk to users this is one thing that we do a lot is is speak to a lot of our users and understand what they're trying to build and where they're struggling and and that's what drives our road map. Um there was this

uh microbiologist uh who had this problem for 10 years that she uh whenever she had an audio book and she didn't like the narrator, she couldn't finish the book. So she built this entirely completely new experience of audiobook where she can import a book,

select different uh narrators from 11 Labs uh voice and then export a completely new experience on the platform. There's this jewelry owner um

platform. There's this jewelry owner um and she had this problem where whenever somebody would walk into the store to uh ask for a jewelry repair, she had to flip through manual books to figure out the pricing. She built this AI pricing

the pricing. She built this AI pricing engine where she just takes a photo and and it tells her what the price is. Um

we had this gardener who built this entire SAS app to manage admin user and gardener for job allocation realtime location tracking. Um so I think what

location tracking. Um so I think what what's what we are really enabling today is like uh entrepreneurs who want to start a new idea. Uh that's that's one category that we see a lot and small business owners who want to digitize

their business solve a problem and we think the software like need is very personal like everybody has a very personal need for software and and we today allow them to build that.

>> So like anyone who had a niche to solve a problem one day.

>> Yes. Now is the time.

>> Yeah. We think you know when we launched we thought like our product is going to be largely used by semi-technical people PMs designers but what what we were surprised to see is that all walks of life uh you know and a lot lot of

business people a lot of like first- time entrepreneurs are using us uh today >> yeah today I was talking to some like a couple of people who are in Spain and they are filmmakers and they are using our platform for like last 3 four months

and they've built a because they're arts artsy people they can bring their arts and to it and like the the website that comes out is phenomenal like we we looked at the website and we were like how did they build this right and they

have no technical like coding background and uh yeah pretty amazing to see >> that's cool how often do you speak with users like you have so many you cannot speak to all of them >> uh no we we don't I mean we we generally like I mean we do a lot of support

support ourselves right so I mean whenever we see a supporting >> speaking with these users like the more the most demanding one is probably >> most demanding one something that catches our eye you know and like and people are very kind on on Twitter as >> I remember the first week we launched

like I was just stuck to my computer answering support tickets uh for for the first you know speak and and >> maybe let's speak about launching here because uh I mean you got that crazy goals in just a few months like uh and I

think you started with an invite only period right can you tell us more about how you approached launching?

>> Yeah. So I think we when um when we were about to launch uh just before we we ran like a very small alpha right and uh just to test the product out very small

>> very small like 50 people 100 people initially very initially and >> and um because we knew internally that that product was working and we could we could test it against other uh platforms out there and and so we wanted to just

you know get add things out but one of the things that we realized very early on was that um it's a very crowded market like we are late entry into a crowded space even though we have a superior product like we need to stand

out and and get the word out. Um so I mean just before launch like 4 weeks before launch we had like no growth people in the team zero idea of how to market things and then uh in four weeks we educated ourselves like how do we

sort of you know really really uh you know get the word out and >> tell us more like what did you do?

>> Yeah. So we realized that influencer marketing is is one of the uh you know like in social media is something that is really powerful for for us to get the word out. Um so first we um ran a beta

word out. Um so first we um ran a beta beta program um and partnered with a bunch of influencers and and gave them invite code which helped us track them back in in terms of performance and um >> so people could not sign up by

themselves just yet. Yes, that's through the invite code of the influencer they could get access to.

>> They could access the platform. So we

see so we partnered with them. We first

figured out who are the best influencers. We figured out like like

influencers. We figured out like like what does Tik Tok algorithm want? What

does Twitter algorithm want? What does

Instagram algorithm want? and and

figured out like how do we create content that would resonate with uh their audience, resonate with the algorithms and and finally like when we launched we were able to get uh word out through these influencers initially and

because we have this tracking built in we were able to track like which content is converting better, which influencers are converting better and then once we

open up the public u you know indeper the influencer had you could see like how much of that would convert based on the code.

>> Yes. Yes. And and that sort of that allowed us to initially get I mean most of the time like people spray and pray like and we were just methodical about tracking these data initially.

>> What did that what kind of choices did you do after that?

>> Yeah. So we so we understood which market what kind of content is going to work where and for example like one of the things we realized that if the influencer is building the content themselves for example like most people when when when we partner with them they

um we ask them to build an app themselves on the platform and and and and obviously we get a lot of feedback but but the whole building journey is so exciting and and and and that connection with their audience. So that's what they kind of like share online with their

audience.

>> Right. Right. And for example like you know if we are you know partnering with a gaming a gaming uh influencer they will build a game right and then show it to their audience and that was the initial thing which which got us you know initial initial traction and then

sort of word of mouth you know started to catch up and and and lot of the growth now is like organic.

>> Do you still work with influencers today?

>> We still do because one of the reason that another reason to sort of invest here is that we also want to build a distribution channel for the apps that are getting built on the platform. like

we want to prioritize this and offer it to our uh builders, right? Because when

they have a new idea, >> they have like some kind of like I don't know showcase of apps.

>> We we do have a we haven't done a good job at it. We should uh and we will soon but uh but one of the reasons like strategically to invest in the this marketing thing is that we want to be able to build a marketing channel for the apps that are getting built on the

platform as well. Yeah,

>> that's awesome. By coding tools also attract a lot of uh like first app tourists like people who are going to just do an app and leave like how do you make sure you get a good retention? Uh

so uh I mean we obviously I think there is a lot of appetite to to try out new apps and we see a lot of users come in you know like try something and and and uh move on and but I think what what we

we are able to attract and retain is the serious builders people who really have an idea in their mind like entrepreneurs >> power users >> power users yeah so we we internally call them power users people who generally have longer prompts people who

have a serious idea in their mind and they're able to build more on the platform their attention is is upward of like 85 90% on the platform. So that's

the actual thing you are tracking.

>> So we really care about the power user usage on the platform and they're driving a lot of revenue for us and and and uh and and their attention is pretty high because they're able to build app and >> so the long tail of like like what I

call tourists doesn't really matter to you just focus on the retention of >> Yeah. So I think and and some of these

>> Yeah. So I think and and some of these tourists will also convert like you know but but but largely they focus on people who are like who really really want to launch an idea who have a use case in mind and we see like most of these people are like domain experts like who

have had a problem that they personally resonate with or have seen or and a lot of our users uh tell us that hey I've gone to a dev shop gotten a code for the software I've gotten a code of $100,000 to build this and I've come to emergent

and in under $1,000 I've been able to build this.

>> So they're serious about the app they want to build. They were nearly ready to pay someone to do it and then they realize they could do >> and they tell us that emergency steal for me because you know like >> uh

>> you don't charge enough.

>> That's uh speaking of money you also recently announced a big series with light speeded right. Yes. Like 23

million. Uh when did that happen? Like

did you have already a lot of traction when when you closed that round?

>> Uh it it actually got closed pretty quickly like once we launched. uh

>> like shortly after launch >> short two weeks from the launch like we were able to close that um and we I think at two two and a half three million error when we sort of raise that >> okay so there's like there was like some

strong signal already that things were working really well >> yeah I think people tried our product and and and you know saw that our approach to to problem solving was very different than >> you mean the investors tried >> the investors yes and and uh >> that's good to have a product that

people can sell sir >> I mean that's the best thing like you know like when people tell us we say go try the platform and you know so >> and you can also try the same prompted on multiple platforms and you you can easily tell, right? So, which one is

working better?

>> So, they're they're feeling like their appreciation of the product was pretty high.

>> That's what you to raise easily.

>> That's that's awesome. Uh what are you going to do with the money?

>> Uh so, a lot of the investment is going to go in just building a world-class team. We already have, you know, a

team. We already have, you know, a really really talent dense team, 12 engineers right now. And and we're just 12 engineers. 12 engineers and you're

12 engineers. 12 engineers and you're like over 15 million. That's pretty good numbers.

>> Uh yeah and but we are now expanding the team like we we want to bring in top talent and and also a lot of the money is going to go into research that we want to really focus on building the worldass agent and of course on growing

the platform. So maybe there's a message

the platform. So maybe there's a message here like he may want.

>> Yeah, I mean we are hiring uh we just opened an office in SF. Would love to you know meet people uh and we're hiring people in post training.

>> Good engineers, researchers, >> engineers, researchers, marketing people, anybody who's creative in storytelling would love to connect.

>> That's awesome. Good to know. Uh so

what's next for the product? Like what

do you think you're going to be able to do in the next six months or >> I mean there are a lot of things which are like already launched but we haven't like marketed it well. So for example, mobile is already out there. You can

build mobile apps on our platform. uh

it's kind kind of like a hidden secret on the platform because uh the mobile app that comes out of the platform also has a back end associated with it. So

it's like functional out of the box. Um

there are like multiple other things happening behind the scenes like custom agents.

>> Yeah, we we have ability to build custom agents on the platform. So you can come and and and build uh agents that can do do your workflow. Uh and a lot of our investment is going to go into just perfecting the uh software engineering

life cycle of the product. Like our core thesis is that uh I mean we look at our products just like if we were to prioritize the best engineering team in the world like what would that look like and and still we have a lot of loose pieces to to tie up. So a lot of our focus is on just quality of the platform

going up.

>> That's cool. So that's emergent. What do

you think about the the market itself?

Like where do you think we're going in the next like two years uh I mean a year ago none of that existed right?

>> Yes. I mean I I mean I feel like there'll be a billion builders uh in in next couple of years and we'll see this new explosion of ideas on on the world.

Great for YC because you know a lot of new entrepreneurs are going to come online and and and build great things and and I I think everybody has this desire they have an idea in their mind that they want to explore and and I

think truly platform like ours enable them to uh you know break out.

>> So any like parting advice for founders or aspiring founders? I mean firstly like try emergent because I think emergent could really help you accelerate your uh building from YC trying emergent right now. So like

building their startups on on emergent.

>> Yeah. But I think in general I would say like I think AI is a big reset right. I

feel that like um it's going to be the next 20 years is going to be about AI.

It's just we're just getting started. I

tell people that this is Bitcoin and dollar one right now. And so you should just be bold and and you know uh find a problem that you really like and just go go all in.

>> That's awesome. Mukun Madav, it was great to have you with us today. Thanks

for sharing your story.

>> Yeah, great to be here. Thank you.

[Music] [Music]