Michael Bajwa, CEO and Co-founder of Zaro: BREAKING: Zaro just raised $5.1m pre-seed

0:00Hello and welcome back to Scaling Europe show. I'm Seb Johnson. Today I've got Michael on the show. Michael is a co-founder of Zaro, an amazing tool with some amazing news today. Take it away.

0:10Michael Bajwa:Hey Seb, great to be here, mate. Yeah, so co-founder and CEO here at Zaro. We just closed out our$5.1 million pre-seed round led by Cherry Ventures. So yeah, we had a couple of angels that kicked on as well. So we've got Thomas Domke from GitHub. He's the ex-CEO there. Thomas Wolfe from Hugging Face and that's pretty special to us because basically what we're building here at Zaro is a memory layer for agents so what we can do is pretty incredible we've built this context layer and effectively what this allows agents to do is have extreme long horizon task and so a couple other angels on here so we came from Convergence I was very fortunate to be the first hire there with my co-founder Chen Zhang being the head of engineering.

1:01Michael Bajwa:We were acquired by Salesforce 11 months after getting the capital in our accounts. And then, yeah, we both left to go and start Zaro. So we're a 13 strong team. We've had the capital for about 10 weeks and yeah, we're just getting started. Amazing. Look, it's a great roster of investors. I want to talk about the conversion story because I think it's such a great story. Before we do this, let's talk about the round. I love the Cherry Ventures team. They're a great team. What gave them conviction to come on board? Yeah. So Dinika Matani was our partner there that made the investment. And from Cherry Ventures to that, I mean, they're an incredible VC.

1:42Michael Bajwa:What really came out was that they moved aggressively. They felt like a US firm, which we really liked. And effectively, the thesis there was that we haven't had this chat GPT moment for enterprises yet. there hasn't been this thing that is stuck and sort of like proliferated across an entire enterprise stack and i think what was uh interesting to them is that we had an agent company that was sold into a large enterprise like salesforce so we had that hyper growth hyperscale zero to one type moment but then also we saw from the inside some of the structural disadvantages that you can have at a larger sort of legacy sas tech firm like salesforce um so marrying those two things together was like a pretty obvious choice and then my co-founder chen uh zang he's you know generational talent one of the best infra ai engineers i've ever met um so very fortunate to be building with him and um he's had experience across like poly ai and a bunch of other ai firms as well so it was a it was a very uh very easy decision to leave our salesforce to go off and build with chen and how important was it for the investors the experience that you both had at Convergence?

2:52Michael Bajwa:Yeah, I think the main thing there was the pace, right? So from zero to one, and then also being the first hire there. So I was very fortunate to be in some of those early IC rounds. So I was there pre-product, pre-timesheet, pre-deck, and then receiving the funding and then just like going out with a bang. And then Chen came in as a head of engineering. So head of product, head of engineering, like we have this like really good synergy like this great working relationship and so we'd work with each other before and then through that we met some incredible engineers and designers um and and through that story right we went from like zero to one mil around about 10 weeks after launch and we directly competed with operator and h company and uh we we absolutely you know we did very well there uh so i think one of the first companies to have a multi-agent system in production it was effectively a web agent that would go out and complete task on a user's behalf but frankly the web agent's accuracy was relatively low and basically it would lose intent in detail as it would try to complete a task so the context window would fill up and it would start to hallucinate off and do sort of strange things then we got acquired by salesforce like hey these guys asian force they have all the capital in the world so like they should be able to solve this context problem with context windows filling out for agents.

4:20Michael Bajwa:But then we got there, we realized that when you're building on legacy databases, they're not accustomed to this new agentic era. So structurally, trying to patch in agents on legacy systems is really difficult. So once we saw that angle, Chen and I were chatting, we're like, hey, we should go off and build a company around solving this problem. Amazing. And, you know, for those listening, I've been using the product and it's absolutely crazy. Like for me, I think I said this to you, but it was like the lovable moment almost, you know? It's like the first time you build something unlovable, you're like, wow, this is, like it's, you just, you see it appear before your eyes.

5:02Michael Bajwa:Now, yeah, you know, very simply prompting and connecting tools has just made it accessible in a way that, For me, it wasn't that accessible. The idea of being able to automate workflows, use agents, both in the browser and connected to all my systems, felt like it was deeply technical. And for me, at first, this is a really game-changing moment. I think there's going to be other tools like yours, but we're going to see a huge influx, I think, of tools like this that make the technology accessible to people. In a way that Lovable has made building websites accessible. this agentic this level of agentic technology and automation has previously been sort of withheld to quite technical people and for me it's just been crazy like it's been absolutely insane to be able to do the things that i've done in like no time at all yeah it's pretty special right um i had a very similar feeling right so when we have that context layer right to describe infrastructure to someone on a whiteboard it's kind of difficult to say that hey now your agents are much more accurate and you can do long horizon task to me as someone who like works in product and growth right um that to me is not really a value prop unless you are deeply deeply technical right yeah and so what we wanted to do especially is make this accessible to anyone that is non-technical and make it so easy to get sort of frontier capabilities without needing to set up clawed code, connecting it to 12 different MCPs, managing context across all these different silos.

6:38Michael Bajwa:You have one tool with your context layer, the ability to build apps from the data that you have, and then you patch agents into that. And you can set up things like, you know, for example, your whole scaling studio that now can manage all your inbound, find funding announcements. That software that you've built for yourself, that doesn't exist anywhere else, but it is particular to your use case. So when you can generate applications and agents just from the data that you have, then you have these very custom sort of applications and agents just for you. Now, the special thing here is that structurally, if I'm a tech company, I want to host and keep all of your data in my system, right?

7:20Michael Bajwa:That is a structural vendor advantage that's come back from 25 years ago, right? You have a deeply verticalized SaaS tool that keeps your data in there and then you continue to charge rent. Now we want to give power back to the people. So now you said you own that data. That is yours forever. That app you build, those agents you build, they're yours, right? They're not going anywhere. And so when you shift that across, you get all these interesting advantages. And number one, you can get like very high quality apps just from a prompt. And then you can set those agents up with just a prompt as well.

7:55Michael Bajwa:So you don't need to use like node-based tools like N8N. It's all from the Zara system, right? And then secondly, as well, is that we're model agnostic. So if I'm a hyperscaler or like if I'm a large model company, I create a very expensive hotel, right? I create a frontier hotel and I need to pay down the capex of that by charging up frontier model rates to recover those costs. And my goal is to keep you in our system, right? because i need to recover all those costs i don't want you to use any other models and that's exactly what thomas wolf saw with our investment right because he owns a company that is specifically designed to have open source models as a platform right and then that whole context layer right that's effectively our our github or like git infrastructure that thomas domke saw from uh github right he was like oh that makes complete sense so we saw that from the angels right they got the thesis and um now what we offer is like insane frontier capability for anyone to build on regardless of your technical ability yeah it's amazing and it's just uh it's been a huge unlock for me in the business but i guess like today you're officially coming out of stealth so like what what's been going on behind the kind of closed doors what kind of stage or product development go to market can you talk a bit about that yeah yeah absolutely Yeah, so it's an exciting day for us.

9:19Michael Bajwa:We've got an absolutely crack team. So including my co-founder and myself, we've got 13 people so far in 10 weeks that just saw the mission and they jumped on board. We have five ex-convergence people that have come across to us, including our designer, a couple of engineers, Chen and myself. We have the ex-growth advisor for Lovable that took them from zero to 100 mil ARR. Wow. And we also have FDEs, cracked operations people. We also have people that are running our SLG motion. And they're really important because what we're doing in parallel, because people like you that have their own businesses, you're a founder, you have all this data, but you want to connect all your tools.

10:03Michael Bajwa:That's our PLG motion, right? So founders, operators, small teams, they can see the value of Zaro sort of immediately, right? But then you have these enterprises at the same time that are struggling to get agents to work for them. And that's our SLG motion. So what we're doing is like, we're doing both at the same time, right? Which is very aggressive, but we think it's right because what we'd like is our PLG motion. So people that see the value of the product immediately, bring that into small teams, and then we're going to convert them to enterprise contracts. The approach is pretty aggressive because effectively we're asking, we're telling people that, you know, now you can get these insane new capabilities but you get to own your own context as well and that's not going anywhere and also where it's almost 10 times cheaper than just running a frontier model we have a very intelligent orchestration that selects the highest performance model for the cheapest cost for the users but if you are an enterprise with your own sort of llms that you're already paying for we can patch them and you can bring your own key amazing it's um yeah it's great um and like I guess you mentioned NA10, you mentioned something like, where do you see yourself sitting in this world?

11:13Is it a new market of sort of like opening up automation to a general public? Or where do you see yourself sitting in this stack between NA10 and the models? And yeah.

11:23Michael Bajwa:Yeah, it's very interesting. I almost feel like it's a new category of product. Yeah. Right. So you have companies like Love of War, which create these beautiful websites for people. and they're very good for internal prototyping, but not connected to your data, right? Or you can't build agents on top. And then you would have a platform like Notion, right? That holds your context, but at the same time, they own your context, right? That's theirs. You're paying your subscription fee for Notion, right? And at the same time, you have companies like Victor that are coming out and saying like, hey, we have our Slack bot, our Slack agents, and they can generate pieces of collateral.

12:01Michael Bajwa:It's kind of the unison of three things together. so what you have is a context layer which is yours that's where your data lives but the way we've architected that allows you to build these like cutting edge applications on top with just a few prompts and then you can make agents read from the application all the context that you have to go off and do like whatever you'd like so so for example in your instance right like you've never now missed a funding post because you have these agents on x and you know other platforms and they're always scraping um the web and updating your studio that you have right and at the same time like you're getting all this in all these inbounds from founders that are trying to get onto your podcast by the way thanks for having me on uh and uh and uh you're trying to manage all that as well right so now you have this like operational machine that runs your business for you and you just use natural language to set that up yeah and that's the thing that's been really powerful is being able to have the app and the agents and still have the chat to be like hey actually uh like run it again you know like i've got automated things that like tell me every day like an email being like hey here are the big stories from this morning that you may have missed uh here here i've got a file which i've got my press releases on but i can just be like hey what's going on today like what press releases are going out today that i think that you think i've missed and it'll just do it and i think that that having that back and forth is also is also really interesting and let's go back to convergence right because you were on this rocket ship 11 months you scaled revenue really quickly what are some of the lessons from that journey that like going from first operator to founder you are now bringing with you on your journey yeah it was a wild ride uh so at the start it was uh myself and the two co-founders uh so it was very interesting to see that whole sort of like raise process all the way to like hiring the first team, scaling that team.

13:53Michael Bajwa:And then I moved into the growth role when we went live and we went directly up against operator. I mean, that whole process took 11 months, but when we were in our IC with Cherry, I told them, I was like, you know, I'm like, that is, that is not the goal, right? Let's like, let's beat that. Right. So if I, if we can do that in 12 months, why can't we do it in three months? Right. Why not? Why can't we accelerate this? and already we have 13 people with us now in 10 weeks from the money hitting our accounts we have our plg motion all set up we have our slg motion set up as of the recording of this we have eight design partners already and that's in a matter of 10 weeks right so we learned how to ship at speed and then not let other things slip right so i think that founders today like either you know they're too technical or too operational right but what you're still building here is a business right and so for a business to run at lightning speed your ops need to be covered right if i need to set up an entity in a different country if i need to set up the options in the emi if i need to set up an office if i need to get contracts all signed and counting legal that infrastructure has to be set up but what's so cool about zara is that we've 10x our ops people as a result.

15:13Michael Bajwa:It feels like we have an army of ops just looking after us. And then from get-go, you set up your PLG motion and you already start your design partnership conversations right from the get-go. That's not something that happens later, that happens when you get the capital, right? You are testing and iterating. So I think already we've had over 100 user interviews. We've had five, 600 beta users on a platform that we're constantly iterating on. And we've had 30 experts and that's from uh james lowe's ethos right that have come onto our platform and built out expert workflows for us that's already going live so when you we've seen it once before the question is how do you compress that in to you know the shortest period of time possible and and what does that what would you be happy with you know like given that your benchmark your baseline was like you know a million ari in 10 weeks a convergence do you have a view of like okay where do we want to be in three months six months a year you know you're obviously very ambitious you're shooting for the moon you know what does like success in year one or six months look like i think we're going to have more people that realize that owning their own context is extremely valuable right as token cost and model costs increase as the large providers are trying to make a return on their large capex investments and training these models you're going to increase prices right and so people are going to start to realize that this stuff is becoming very expensive and owning your own context and getting insane capabilities from that is going to be extremely important right if the larger companies can't do this because their databases aren't structured in a way that allows agents to retrieve right and do amazing things from this it'll be companies like ours that come out and show people that what they're building like you can build things just for yourself and for your org right so seb for example like your scaling studio let's say you want to scale the team everyone in your org will have access to that scaling studio that you've created they'll be able to retrieve from that and set up with the permissions that you would like right so now now internally we haven't had to buy any hr software like um operation software finance software we have everything patched into zaro and our sas spend is extremely low as a result because we've built everything internally very quickly with just a couple of prompts.

17:33Michael Bajwa:And so I think that we're looking for sort of a category defining company here. We saw the very fast acquisition there once. Chen and I have already been through that, but we're looking to just take this as far as we can. And how are you thinking about talent? You know, you obviously had an amazing team at Conversions. You built, I think you said, 13 people. when you look forward like who do you need in your team to to make this successful yeah definitely at this stage we just need the dogs right we need people that are just battle-tested and hardened right and that only comes from being in the gulag of a startup for years and years on end right and so you know i for for us like it's trying to identify that talent like one of our guys runs like ultra marathons right for example we have another one that is like you know you know escaped a war right we have many people that are um ex-founders so over half the team are ex-founders uh we have had many people that have gone through an acquisition uh we've had um people that are experts in their field already so we have um yanis who is our uh our guy that does our plg motion he was a growth advisor for level wall i took that from zero to 100 mil right so i think as well building out of london is a very special time and place uh it's you know the most capitalized um country in europe right now for ai talent we have a lot of major labs coming through here as well um so yeah talent talent is quite ripe and rich here and you know like for example the the the work ethic here is just insane right and i think coming from startups is where we're trying to target from because people understand what's at stake a lot of ex-founders here as well so anyone that is like been through the punishment of a startup is probably right for us right now yeah someone who knows the grind knows what it's like uh can yeah get get in the trenches well look michael best of luck it has been amazing to chat i'm so happy for the team it's been truly a game-changing product for me and i just don't see a world in which it's not successful for you guys.

19:46It's like the time it saved me, the value it's given me is huge. And so I'm grateful for you and the team for helping me out with it. And it's an amazing product.

19:55Michael Bajwa:I appreciate you said. Thank you so much, mate.