Paradigm AI vs. Excel: Building an AI-Native Spreadsheet Startup | Anna Monako, DVC

We should be able to have like a 10 person $1 billion company.

You feel that you don't work for a corporation. You work for a team of like-minded people.

You already on Forbes 30 at the furry,

but it definitely was not the experience that I was like, "Oh, I want to spend the rest of my life doing this."

I think apps like that are super important for freedom of speech.

Schedule these customer demos with no product at all. You can do a lot more than you think.

Hello everybody. Welcome to Deconstructed VC. It's a podcast where Marina and

Nick David

we are trying to deconstruct the way businesses and startups specifically are built. Talk to the founders, learn their stories and you know

and get some inspiration, right?

Yep. Good. Okay. Done with the intro. today we have Anna Monaco who is the founder of an incredible company called Paradigm. is Paradigm the name of the company or just the product?

The name of the company. Yes.

The name of the company.

Yes.

And Paradigm is a company we fought very hard to even get in because it's a very hot and overs subscribed startup. and I I think today you'll understand why.

Yeah. Thank you. Thank you guys for having me. Really excited. But let's talk a little before the paradigm because your story is very inspiring. by the age of 22 you did so much that most of the people still trying to be at least somewhere there. you already on Forbes 30 at the F and u for our fun is very important. You are also a repeat founder right? you already build a company, sold the company and you started doing this pretty early.

how and why? I think the most important the question why.

Yes. Yeah. So, I guess going all the way back, I grew up in New York and I feel like I was surrounded by much more of a finance influence than tech. But from a very young age, I was very obsessed with building things. I would like ask my parents for like things like a micros a microscope or like a circuit builder for Christmas. and yeah, basically started just getting really into like programming when I was very young. all things related to technology. I went to Penn and I studied computer science. And at the time, so I had this like very pivotal experience. It was my senior year of high school and I started doing hackathons and I won a hackathon and the prize was a tour of Google in New York.

Oh wow. And

but what what did you do for the hackathon?

it was it was just a random project. I think we built I was really into building games at the time. So, we built a game to it was an educational game for this one that taught people about like solar solar power, but we made custom graphics.

Oh,

so I did

in a hackathon.

Yeah, I did a bunch of things with like Unity and Unreal Engine. I had many different like rabbit holes in high school. but yeah, I won this hackathon and the prize was a tour of Google and I went to Google and I was like, "This is the best place ever. like I must work here. so my freshman year internship at Penn was at Google. I was like very excited about it. Studied very hard for the interviews. and I got there and I had an amazing experience. But it definitely was not the experience that I was like, "Oh, I want to spend the rest of my life doing this." but I did meet a very important person there named June who is the CTO of Paradigm now. and we basically spent like that entire summer just like talking about startups, loosely ideulating things. and I wanted to be in an environment where I could move very fast.

And I didn't find that in big tech. So I started building things on the side. I worked on two different things while I was at Penn. the first was this app called Ripple, which was this like Bluetoothbased messaging system. where you could open your phone and see around you all of the people nearby and like message them. And it was mainly aimed at conferences and events particularly around when there was it was COVID right after CO and everyone had masks on. So it was very hard to actually figure out the exact people that you wanted to talk to. and yeah, I worked on this with some like very smart pen people who also are still like in the startup world now. and the other thing I did was I started kind of like an agency where I started just doing a lot of contracting work for other startups.

So, I built a bunch of websites for other startups, built a bunch of MVPs, and yeah, I scaled that out pretty profitably as well. And I think the experiences that I learned doing my own things in college was vastly more than I learned in any other experience. And I think a big thing that I wanted to do was figure out a way to actually like do what I love and push myself to the limits of what's possible. And for me that has always been start startups. Yeah.

Cool. So what happened to the company?

Yeah. So I guess we were in a lot of discussions to see if we could like acquire the tech of the company. didn't end up actually going through with the acquisition, but we got like a ton of users. The app is still out on the app store. and yeah, the other company, it was just completely bootstrapped and profitable the entire way and still working with a lot of the designers that

I think apps like that are super important for freedom of speech. like when there were protests in Belarus, we tried to organize something like that, but we figured that is pretty hard to to code something like this. So, you did a great job at making it work.

Yeah, it was very deeply technical. A lot of the Bluetooth stuff we did which was a great learning experience of really being on the cutting edge of what's possible with a certain technology.

And by the way, am I right? You taught yourself programming in in school, right?

Yes. So, in high school, we didn't really have that many programming classes. We had a couple like loosely creative programming classes, but most of the things I did were just completely on my own, self-taught. I really was addicted to the demo, I guess, like the point at which you had been imagining this thing, you can finally show it to other people and watch them use it, which is I think why I was so drawn to like design and visual programming early on. so I actually, yeah, started off built a bunch of games, built a bunch of front-end interfaces, websites, and yeah, seeing people's reaction and seeing them use it was always like an incredible.

What do you think about vibe coding now?

I think that it is amazing in the sense that it makes coding so much more accessible to non-technical people. I think that it is a tool that you have to understand how to use because I also think that with vibe coding like if you are just going to put in a prompt and let AI just like run infinitely and build your entire program it might be able to get to a demoable state but there's like an increasingly big gap between demo and like production grade ready. and I think that vibe coding is like an amazing tool. I feel like I leverage like AI coding tools a lot in my workflow, but I also think it's important that like the next generation of programmers actually takes the time to, you know, read documentation, go deep, understand the code they're writing. some things really can't be vibe coded in a sustainable way.

so I think at paradigm we really try to think of the balance of like okay what can be coded with AI and what do we have to actually like go deep into the documentation the traditional way and apply data structures scalably.

Do you feel it could be a gateway drug into actually learning how to code? I think definitely like I I would say the amount of like satisfaction that you feel from just seeing things work is so much and I think a lot of my flow now when I'm coding features is okay let's quickly prototype let's quickly get something together which does involve a lot of AI and then going through making sure everything is sustainable and scalable. and I think that the more people that are creating things, the more strong like the overall impact of technology on society will be.

I still remember the reaction of our son when he was six years old and he first time caught it in Java first object in Minecraft and he realized that now it's in a game and he his eyes was showing I am a god

like creator of worlds

creator And that's very powerful.

It's very powerful. Yeah, it's it's really awesome.

how did you come up with an idea for Paradigm?

So, Paradigm, I was really obsessed with this concept of agent interfaces. So, this was around like the very very beginning of like the word agent becoming a thing. and I was like trying out every single agent demo online, like using it, breaking it, emailing people, like all of this stuff. And the very common thread was like most agent interfaces were chatbased. and chat is like undeniably a very very powerful interface. But I couldn't stop thinking about okay, what other interfaces are there for agents? And particularly along that thread, I had been using this pattern of using spreadsheets as like CRM or as places to store data over time. And I noticed that this was like super common. And I did this in like my last businesses. All of my friends did it. We had all of these like very important lists of things in just normal spreadsheets.

and I couldn't stop thinking about, okay, like what would it look like if we took this spreadsheet interface, you know, a billion people already use spreadsheets, it's the most popular IDE in the world, and made it AI native. and I think what we realized is that so many workflows are comprised of this like three-step process of like find or upload some information, enrich this information with more specific granular data, and then take action on it. And that is the interface for that that works perfectly is a spreadsheet. You know, you can see all of these cells in front of you and you already know how to interact with them. So I think it was just this obsession with interfaces and this pattern that I noticed of how I use spreadsheets in the past.

I think this is what agentic businesses should be because you know chat is just a common line and it well from one point it gives you unlimited creativity. you can do whatever you want but you have to know what are the options. So building interfaces on top of that like you know cursor is wildly successful because I think they nailed the interface

but they kind of squished between you know the ID and and the models now.

I agree. Yeah. And it's about like taking an interface like with cursor, it's an IDE. People are already familiar with this and how do we transform small parts of the user experience so that they really feel like AI native.

I think there are more spreadsheet people in the world than there are ID people. There are like what 40 million developers in the world.

Yeah.

And I think about a billion people who use spreadsheets. Oh, I know so many investors who still run their VC funds on spreadsheet and not on AI spreadsheet, just general simple spreadsheet.

so tell us how you started it.

Yeah, so I guess I started the company a little over a year ago now. and the initial phases were really just like this rapid iteration and feedback loop. So at the very beginning it was literally just me and June. we would code in this way. So we once we like had kind of validated this main vision of the product, we were pretty much like laser focused on getting it to demo to demo stage because the infrastructure for this is non-trivial. You can't you actually cannot like vibe code this in a day. so basically what we did is we kept scheduling demos before we were ready to actually demo it. So we like had the idea and we would like randomly get in touch with a bunch of customers that needed a solution and we would sit like right next to each other like computers like the closer our computers were we would work on the same branch and just like stream of consciousness just create the entire product from the ground up basically and we'd be like oh no like we have a demo the next day we would get on a call it would like not work at all we would still schedule the next demo and the demo.

Building infrastructure for this was not trivial and we basically actually the first money we took was Y Combinator. So we were in this pressure cooker of getting something to work as quickly as possible. so what we would do is we would schedule these customer demos with no product at all really and we would just stream of consciousness. June and I would sit right next to each other and code until the demo it would like barely work, not work at all, not work again. And then finally like we had something that was consistently working very well. And yeah, from that point on we had we started preparing for a bigger launch and getting the product out there.

But while in YC that still wasn't a spreadsheet, was it?

It wasn't. It wasn't. It was an agentic system like general purpose that worked

and I think we looked at that and that was impressive.

But then there were like 200 companies doing something the same.

Yeah. I think for me the shut up and take my money moment was when you showed the spreadsheet and you I think it was a list of developers and you added a field where the prompt was like go to their GitHub and see how many commits that are AI related they they they did in the last month.

Yeah.

And then you just dragged the the row and I saw it started working and I was like wow. So they made a infra that is scalable to how many rows there are to run lots of agents simultaneously and just do this in real time. That was like wow this is very hard to do

and what was the most impressive thing you did it like just maybe couple of months from the previous product and just two of you so the speed of yeah interactions and like trying that was blowing moment for us.

Yeah absolutely. Yeah, it was it was pretty it was pretty awesome. I think in terms of the infrastructure, we basically built the entire infrastructure from the ground up. we couldn't even use like realistically use Python for our infrastructure. We had to use Golang, which is like particularly a language that's very good at concurrent tasks. and we had to create this like Q system where we would load balance all of the calls like potentially across multiple models, ensure we're not hitting any rate limits. because we wanted it to be extremely fast, like the fastest product out there, also being the most accurate and very cost-effective. so yeah, it was it was really amazing like watching it come together, but I think it was just like pure obsession in the early days. like we couldn't stop working on it if we wanted to.

And and back then you you were still two of you and then we started hiring people and you all lived together in one house, right?

Yes. so yeah, our founding team first was just me and June and then we hired like our first couple of people and we were already we didn't really know people in SF. Our entire team for some reason came from the east coast to the Bay Area to be close to customers to be like on the bleeding edge of technology and kind of the natural thing was we were going to be working pretty much non-stop. So we decided to live together and I think it actually built a pretty awesome culture around our early team in terms of like how bought in we were how synchronized we were and even now we still have a house. and yeah, we're we've started hiring more and growing the team. And not everyone lives there anymore, but we still like all come work in a house every day. And something just feels so special about working in a house versus an office.

It really is like

it's the ultimate startup vibe.

Yeah. Exactly. Yeah.

It's like you you feel that you don't work for a corporation, you work for a team of like-minded people, and you know, there's a lot of agency. There's a lot of ownership in that. Exactly. Yeah.

Yeah. I'm just wondering so most of the most of the product you built just three of you and now just recently you grow to team of 10.

Yes.

what's next?

Yes.

Everybody still lives together or you know an office?

Yeah. we have we have like some of the core founding team still living together. and then some people commute in and we basically use the house as an office. but yeah, hiring was it was really hard because we had very high standards for kind of who would join the early team. and we spent some time just figuring out okay like the amazing thing now is that small teams are able to command like so much more resources. So we actually don't need that many more people in the future to build this like massive vision that we have. which means that each person needs to be the perfect fit because you know we only have like the way that June and I see it is we should be able to have like a 10 person $1 billion company you know so each of those slots are very very important

now maybe a 100 people for 10 bill

you know maybe even less yeah

that would be good

it's so there's always this moment when I think 200 is the magic number after which every startup becomes a mess.

It's when people you hired are hiring their people.

It starts because founders usually hire not only based on skills not only you know a cog wheel to work in the in the machine but also people who are multifaceted and vibe well with the culture. But when they start hiring a lot of times they would hire somebody who would solve the specific function and a lot of times that just breaks the culture and also early employees we see a lot of that in companies that scale to like 200 people very quickly and that was the same problem in Google early Oracle. nobody has really avoided it yet. it's your early employees feel displaced because they used to be you know live in the same house do everything and then like now somebody has a VP title and they have to go and speak to them and they're not regarded as you know equals that that is usually a very hard problem to solve.

So it is really good that you're staying very lean and low population.

Yeah for sure. We built a product that we believe can scale exponentially regardless of headcount. like we don't need to, you know, when we sign an enterprise contract, we don't need to put like three engineers on that contract. The product ideally should be able to speak for itself

there. Do they still require any integrations? Do they want, you know, to be able to populate fields with their internal data? What what is the like general scope that you have to do for a client? Definitely. I think actually all of our customers up to this point, we haven't built any custom integrations so far, but definitely there's a lot of demand for using custom data and having custom integrations. So this is something that we're exploring in the future, but we believe that we can get very far with just the base product and the integrations that we already internally have. So we want to kind of push those more custom deployments down the line.

Mhm. because that's well I would use a feature that would allow me to slice my data

with natural language and but shifting gears some of your competitors being you know air table and Google sheets are trying to go in different directions and while Google is just adding an akai or whatever they call it like into formulas I don't think it's anywhere comparable to what paradigm allows because it's just you know an LLM output that is not even it cannot search online it cannot do things but air tableable is going in a complete new direction they're allowing people to build apps using their air table bases as as back end what do you think about this and do you think there might be future for paradigm and

yeah I think a lot of companies are converging on this kind of app builder idea I think that it makes a lot of sense broadly. However, I do think that there's a lot of sometimes you don't need a whole new interface to do a task very very well. and I think that's one thing that we're seeing with these app builders like a lot of it is oh let's create the perfect interface for this task. But I think a lot of what's really powerful is you can build an app through workflows but also through like familiar interfaces. So I think the way we see it with paradigm is it's not really an app builder in the traditional sense but you can build workflows and you can build apps but we think that this core interface of a spreadsheet is so good and so powerful and so ubiquitous that it can still stay at the center of these like apps or workflows or buildouts that people do are

how autonomous are the agents in in paradigm can they just update things regularly? Yeah, I think it's it's pretty autonomous. So, we have this feature called like web hooks that can automatically like update data over time based on things that happen externally. and yeah, we also are adding a feature very soon where you can like schedule a refresh on data. A lot of what we're seeing on traditional databases is data gets stale very quickly and a lot of people want to like go through check every single field in their database. Is this still relevant? Is this still recent? So that's one thing that we're working on right now as well.

Do you see more users coming from a specific industry like finance?

We actually are still very horizontal. I would say there's two main buckets. The first is finance. So we're seeing a lot in like consulting and private equity where you're creating market maps or creating competitive analysis and they already have like entire teams of consultants or analysts that are making these spreadsheets literally all day within a company. So it's a huge huge pain point there. another one we see is just an operator. So like you're spending a lot of time doing either sales or recruiting and you want to generate like a list of leads, you want to enrich a list of engineers. you find a huge data set that you want to pull specific information from. It's kind of a great generalist tool for operators.

That is really good. I I tried using it in in in a variety of ways and I think what I found work works best for me is that like whenever I do like what you know OpenAI or Google or Perplexity does with deep research when I do my own deep research I usually have a certain number of questions that I want to ask and but then I want as many examples as as I can find.

Yeah. So what I I do is just like write different small prompts to just ask questions and and then populate it with with a list of companies and candidates I want to look at and just have paradigm populate everything. So it's just a very

yeah that's really good

predictable way for me get really predictable results like populate forms and what is the wildest use case people done that you've seen so far? There are some crazy use cases that I don't even fully understand. Like some of my friends are in bio and they're working on like a bunch of biotech startups and they're actually using it to like do specific details of analyzing like DNA sequences. They're like putting the proteins in a column and like every single column is like a property about that protein that they're extracting from the internet and with an analysis. yeah, it's it's pretty cool.

That must be very valuable.

Yes, exactly. and there's there's a lot of these like, you know, like random deep tech related things where you just need to do like a bunch of very like heavy research through like high throughput search on a big database and that's perfect for paradigm.

How do you see your go to market strategy? Like are you going to focus on just enterprise sales?

yeah, so right now we're very focused on being kind of a bottoms up productled growth company. So we plan to pretty soon like launch the product completely like generally available and we actually started doing this with our last launch because we launched a waitlist and we had almost 10,000 people just sign up for the waitlist. So we started letting batches and batches of people off the waitlist and what we saw was that there were great use cases both for individuals and small teams that wanted to use paradigm here and there in their day-to-day and also individuals from enterprises would bring it up and say hey we want an enterprise license we have this like very big important use case. so our strategy to go enterprise in the future is actually to be like bottoms up and product led for now. How do you figure out pricing for something like this?

There there there's no comparables.

Yes. So, basically, we try to have our pricing be as accessible as possible. So, we have plans starting at just like $20 a month, $40 per user per month, like very low cost, even like premium trial plans. And what we do is we give you a baseline amount of usage for that. and it's pretty generous. You can get a lot done, but a lot of users need even more than the baseline of usage. So, we do a lot of usage based pricing on top of that. And the great thing about usage based pricing is you are extracting value for every single thing that you do on the product. And it's not like you're paying a fixed amount and you may or may not use it this month. so yeah we do a basis in like seatbased userbased pricing and then if you want to use it a lot more and be a power user there's more on top.

Let us think do you want to cover anything specific for the launch?

I guess I can just like you could ask like it could be from I guess it's going to release when we've already launched, right?

probably at the same time. Yeah, we'll

try as soon as possible. But yeah.

Okay, cool. So maybe you can just ask me like so like how do you try it out and then I can kind of just like that's that's a good question.

So the launch is probably you're seeing this as the launch is happening. how do how do people try it?

Yeah, so basically you just go to our website. It's paradigmai.com and you can for free sign up, create an account, invite your entire team and try out some of our really cool templates. They're on the website. We have a bunch of templates and workflows that are set up for a ton of use cases. So, yeah, it's available now to try. One tip from me as an early user take a Google sheet that you work a lot with and just completely copy it into paradigm and see the magic

and see the magic. Yeah, like you can add fields and like may you might have been missing data or having outdated data just mask 2 updated. It's going to be incredible.

Yes. Yeah.

Let's talk about something that usually bothered all the founders about the fundraising. you've got amazing investors by your side. You have VIC, you have General Catalyst and founders of a lot of unicorns, intercom, lang chain. what else? Maybe I missed someone.

Dropbox.

Yeah, Dropbox.

Yeah, there's some cool ones.

Yeah. How you did this? I think coming originally to the Bay Area, I actually didn't really know any investors. but I think over time, with YC was a great help. Also, I'm in this accelerator called Prod that was a great help. I think what really helped it happen, like the fund raise go really well, was just having this like genuine obsession with the product that people wanted to talk about and wanted to share. and kind of getting like a lot of my friends involved that are like very successful like AI researchers or founders and just they were like, "Oh, like can I invest a small check?" And then they would like introduce us to a bunch of awesome people. So kind of leveraging your network letting people that really care about the product like you guys as well like kind of join the team and help out has been really great and I think timing the fund raise as well with like nice like bursts and traction is always super helpful for fundraisers.

So not to sit and stuff and like but share your obsession with everyone else. Yeah, I think we we haven't like been all in on building in public, but we've definitely been posting a lot of demos and use cases throughout the way. And now with this launch, we're going to be like fully all in on building in public. Like you will not stop hearing from me and the rest of the team on socials. And I think just showing people what you've built, showing like a real live demo on the internet is so powerful.

And yeah, I think you can be proud of of what you already built.

Thank you. I hope you know users going to start sharing like templates that they create.

Oh yeah

because that's u like a lot a lot of people do some crazy stuff that is very useful for other people.

Yes. Yeah. We actually have a feature where you can share templates with not just with your team but also with like the broader paradigm community. So really excited to see what people come up with with that and it's been cool so far. like I can I can think of I don't know infinite number of use cases like vacation planner hiring assistant I don't know M&A assistant

due diligence

we are using it to screen like a lot of companies from community minute as well.

Oh yeah like just looking at a YC batch in one go the best use case. Yeah, it's great

cuz yeah, nobody, let's be real, nobody nobody now like waits till demo day YC like the moment they started publishing the list of companies in the batch every like every everybody knows everything about these companies like in the next five minutes and if you don't well try using paradigm

exactly

I have one more question usually building company is already stressful but building AI company in a this time in this time in very aggressive market when every everything is changing constantly is even more stressful. How are you coping it? Do you have any advice for founders how to keep same? Yeah, I think like at the core if you're doing something that in itself like despite like certain dynamics of the market just like makes you really proud and like you know I feel like with paradigm I can be like the number one user of my own product. I think just building something that you're so deeply passionate about and falling back on that when there's the these like external forces out of your control is really useful. And I also think that just remembering that it's okay and it's actually good. I think that it's a very competitive space like most things that are really hard and really impactful are very competitive and I think that's that means that it's worth doing and you should push even harder and just know that if there's a lot of competition you know I think there's so much value that's going to be created in application layer AI that genuinely like it's not a zero- sum game many companies will have amazing outcomes and win and just keeping in mind that you know things in very competitive spaces are really worth doing.

Yeah. And I think the more competition there is the more users notice that you can do something the like a new way.

Exactly.

Apple comes tomorrow and says hey like Apple sheets need an a revamp or numbers numbers they call it need a revamp. Do you want to come join with your team and work at Apple? Would you take something like that? Not. I think right now we have a lot that we can do internally. I think at the end of the day what I care most about is Paradigm really like achieving its mission of like allowing individuals to create like an enormous business impact in a very short amount of time. And like if at some point much more down the line that means that like joining forces with another company that shares the same mission is the best way to do that we would consider it. But I think right now we are just like so allin and I think we have like a great team going, great momentum that we're not really thinking about that.

Yeah. Well, maybe you can acquire your numbers from Apple.

Exactly. That's what I was thinking. That's what I was thinking. Any advice for the founders how to like how to choose the the right VC fund?

Yeah, I think going off of references is really useful. so I think both with you guys and with like the other investors that I've let in, it has literally been because my friends have directly called me and been like, we have worked with this person and they are great. And I think there's nothing more reassuring than talking to someone that has already worked with those investors. so yeah, even ask the investors, like if you don't already know people that they've invested in, ask them to connect you with those people. And I think you can never go wrong if you hear it from like firsthand sources.

So reputation is still everything.

And I think it's also a good fundraising tip. If you don't have a long list of investors that want to invest in your company and you're talking to some, still ask them for references. That will at least create an impression that you're over subscribed and choosing.

Yeah. Exactly. Exactly. And I think maybe the last question so because the nature of our podcast is to deconstruct startups and to inspire founders. I'm a founder. I have an idea but I'm frustrated because it's u yeah that's it's it's hard. what would you suggest? Yeah, I guess I would suggest like thinking about like what is your why on like why are you working on this thing? And usually for founders, there's like a really strong answer to this. and I think when you're frustrated, it's important to realize like the potential impact that it could have and the reason why remind yourself of the reason why you're doing it in the first place. And if that why is still really meaningfulful to you, that's a great reason to kind of push on. And I would say like there's always ways where you can change the dynamics, right?

Like with with paradigm, there's been like weeks where it's like, oh my gosh, this is like there's so much going wrong. Like this is crazy. Like this is so frustrating. All of these things. But in reality, like it's in your control to like take a next step that could produce like a very dynamic impact and outcome on the company. So I would say like be ambitious like do crazy things, switch things up, change the momentum. Like you can do a lot more than you think.

Amazing.

That is a good that is a good advice. Thank you so much for joining us on today's podcast and I hope that paradigm becomes that 10 billion company with less than 100 employees while making like half a billion people at least depend on the product love it and and enjoy it.

Awesome. Yeah, I will get back to it. But thank you guys so much for having me. This was awesome. Really appreciate the time.

Cool. Thank you.