Scaling a Unicorn at Machine Speed: Mirko Novakovic, CEO of Dash0, on AI-Native Engineering and Growth

0:00This episode is sponsored by Cpoint, the business account built for startups. If you're a startup founder fed up with finance admin, you need Cpoint. Cpoint is different to neobanks because it connects all your bank accounts and your strike accounts to your cash position, burn and runway instantly. It automates bookkeeping by pulling invoices and receipts from your whole team's inboxes, so you just sync everything to zero and pay outstanding bills with a click. It has a 3.49 % yield on treasury and real human customer support. Find out more at seapoint.co. That's S-E-A-P-O-I-N-T dot co. Use code UNICORN for a free month.

0:35Seapoint Treasury is a money market fund. Rate recorded at 1st of October 2026. Rates are variable and subject to change. Capital at risk. Hello and welcome to another episode of Riding Unicorns. Today we're thrilled to be hosting Mirko Novikovic, CEO of Dash Zero. Dash Zero recently announced a 110 million Series B investment, which is amazing. And also, Mirko, I saw that you posted recently that you'd 4X'd ARR in just six months, which is incredible. So it's great to have you on. So, Mirko, maybe you could explain in your own words what Dash Zero does. Yeah, absolutely. First of all, thanks for having me.

1:14Dash Zero is an observability company. So observability is a category of tools that help customers to find and solve issues with their applications. We provide the right data. And today with AI, we even provide fixes for problems so that they get automatically fixed. Awesome. And so in a world of more and more code being written by AI, how does Dash Zero fit into that as a supporting tool to this very fast-paced environment now that software engineering is? First of all, you need more observability, right? If you create more software and faster, so you release more frequently, you cannot run without a safety net.

1:55So you need observability to understand what's happening. And what we say, because code is now created at machine speed, not any more human speed. You use agents more and more for coding. Also on the production side, so if you look at this software, you have to have machine speed because if you would use still the old way of looking at data, manually figuring out problems, doesn't really work anymore because the change rate is just too high. And the other big topic in our space is the amount of data that's created. So you think about whatever you do when you create a piece of software, that software creates telemetry.

2:28A user logs in, so it creates an event, you get data, you get metrics, a response time, all these things. And the more you have, the more data is created, also by AI. That means that you need different types of data storage systems because we need to store, analyze that data. And if it becomes too much, it doesn't work anymore. Yeah, absolutely. And your seed round was announced at least less than two years ago. And you're now already a unicorn. You are a serial entrepreneur. So I'm sure you've taken a lot of learnings from previous experiences into this. But apart from being a brilliant idea and very on trend with what's happening in AI and engineering, what has been the secret sauce that has helped you grow so quickly and so successfully in such a short period of time i'm a big advocate that there is no secret sauce to be honest i mean building a software company is not that hard so you build it brick by brick i mean you cannot step over one of those bricks so you have to build a product first you have to find product market fit, then you have to create a go-to-market engine, marketing, all this.

3:40But each of those bricks, each of those components, if you look at the fundamentals, are not that hard. I think the hardest part is making sure that you orchestrate it, that you don't forget about it, and that you constantly work on it so that you have a compounding effect. I think that's something I posted recently about that, that compounding is something that most founders don't realize. Because when you start with anything, is it a new website, your new LinkedIn profile, you put the software out, normally nothing happens because nobody knows that you are there. Nobody comes to your website, then five people come to your website, then 10, then 20, then 40.

4:18And that's compounding, right? Over time, once you create content, you create more content, you get found better on Google, you have more followers on LinkedIn, but you have to create that. And you do that by working hard every day and then you just need some time that these compounding effects take place but i think a lot of people give up because it feels like losing when you start i mean every founder has had that moment when you release your product and you are super excited you do that press release and then a week later nobody is showing up and you see your google analytics that there are three people on your website right it just feels like shit but that's totally normal i mean there are a few exceptions, right, where you have that viral effects.

5:01But I would say 99 % of the companies out there, they have to build that traction step by step, brick by brick. And you have to be patient and you have to do the right things step by step and just work hard and do it. But there's no secret sauce. There's no secret sauce in my point of view in marketing or in sales or it's the basics. You need good people and then you need to follow the basic process steps and then you need to repeat it every day, learn, adopt? Market is changing all the time. Yeah, so there's no like overnight moment. And it's about focusing on every part of the business and making it better every day.

5:38And hopefully those inputs combine to create this feeling of momentum, and the numbers start going up. And also, you need to learn every day, right? And be data driven. I mean, let's think about outbounding, right? There was a time where email outbounding worked really well. And I just read some CEO of a very popular AI company that did a lot of email bonding. He posted email is not working anymore. Now you have all the AI tools and you get too many emails. It's so much noise. And for us also, email doesn't work anymore. So you need to switch to phone calls. So you need cold calling. And that can change.

6:13Maybe three years ago, cold calling doesn't work. Use emails. And today it's cold calling again, right? And maybe tomorrow it's something else. So you have to be always listening, what's working, looking at data, testing different channels, and then go from there. Same with product, right? You need to bring features out and see if people are adopting it or not. Yeah. And one of the aims of this show is to show people what success looks like in the venture capital market. and also educate people on the fact that not every business suits that venture model. Part of that is in the venture world, it's very momentum driven.

6:52You want to move very quickly, blitz scale, and then keep going and keep going. When you find something that's working, how do you lean into that and turn a bit of momentum into a lot of momentum? Do you spend more money, hire more people? How do you go about going, wow, we've got something there that's working. Let's lean into that and really supercharge this. Yeah, that's a very tough question to answer for every founder and CEO, right? Because if you scale too early, you will burn a lot of money and it's frustrating, right? If you scale too late, then you will lose a lot of the momentum, right?

7:32And you cannot really build it up anymore because you do not have enough capacity. I mean, that's a little bit because this is my third company. And every time I did this, I leaned in too late. And so this time I said, okay, this time I will lean in big time when I feel that we have momentum. And I felt that probably after four or five months when we started selling, we went pretty quickly to, I don't know, a million dollar in revenue, like in four or five months and got 100 customers. So I was like, okay, that's good enough, right? And then I stepped on the gas. I hired salespeople, bigger packages at trade shows, and just went all in.

8:10But then you have to have the confidence that you can raise money to finance that. You need the momentum to show up because otherwise you are in a bad spot. In venture capital, that's the game, right? You want high outcomes and everybody knows it's risky and you have to take that risk. If you don't want that risk, then don't do venture capital. But I think if you are in that venture world, it's kind of what you have to do. That's why a lot of Silicon Valley startups are so successful. I'm from Germany, especially Germans tend to be more conservative. And that doesn't create these big outliers.

8:43It's sort of counterintuitive to human nature to sort of go, I'm going to shorten my runway without the promise of the next funding round. But actually, you've got to lean in and go aggressive. And I always say to founders that growth outperforms anything when it comes to fundraising. If you're growing exceptionally quickly, fundraising gets easier. So you'd rather have a slightly shorter runway and amazing growth metrics than try and elongate your runway, but petering towards your targets. A hundred percent. I also did a lot of interest. I also recommend always to founders is just focus on that one metric is revenue growth.

9:21All the other metrics, that's my experience, do not matter if you have high growth. If you have high growth, efficiency, especially at the beginning, does not really matter. That's different if you are C or C, D or whatever, but at the beginning, it doesn't really matter. So you can also grow high inefficiently if you can grow, right? You need to show that growth. But as you said, it's super scary. Also for me, it's always scary to know, oh shit, I have 10 months runway. I'm spending a lot of money. If I don't see the results in two, three months, then your runway is short and you need to raise.

9:57So yeah, it's risky. it's risky but it is the game and then obviously with scale becomes new challenges so at the beginning you're focused on that one metric and you're trying to you know grow and listen to customers and get their feedback and make the product better but as you grow you start getting larger volumes of data larger teams larger complexity tell us about some of the challenges you've experienced with dash zero around particularly on the data volume side because that must get really complex as you scale. Absolutely. So yeah, you need a really great technical team that knows how to handle that data.

10:36But there's also another thing you have to do as a founder and CEO and say no to customers. I mean, we had some deals where you look at probably 10 million plus per year deals, which are so exciting. But at the same time, you know, we cannot handle that at the moment. We are not there yet. We need another year or two to really build that infrastructure. And those customers require you to be up and running all the time, right? You cannot trash your system or not handle the data or drop data because that can kill your company, right? If you take on a too big customer too early, that can be very demanding for your team.

11:15It will slow down your roadmap. And at the end, it can really be the end of the company. So you have to say no, even if it's hard and then grow sustainable. I mean, it's physics. At the end of the day, you cannot handle unlimited data from day one. And you mentioned having an amazing engineering team. Have you got any tips or lessons that you can share around how to hire amazing people? What do you do that you think really helps with that process of building a great team that can deal with the speed, but also the responsibility of the customers and growing quickly? The number one thing is that the first engineers who hire your founding engineers must be exceptional because exceptional engineers create this environment where other exceptional engineers want to work.

12:01That's kind of the starting point. If you don't have that super exceptional engineering team at day zero, it will be hard to get it later. And then you have to keep the standards high. It's like this brick by brick, right? You have to reach out to engineers, especially now we have recruiters and internally who do that. But at the beginning, I reached out to engineers on LinkedIn. We looked for profiles and then we approached them. If you are reaching out as a founder, it's different. So you just have to do it. You have to find the right talent, very high standards, especially for the initial team.

12:31And then good people have friends. They know other good people. So you get a bigger network and that's how it's working. And I think there's some interesting stat about, I think in the US, they hire a head of talent on average, employee number 13. and in Europe it's more like employee number 13 so you know almost what is that 17 people later companies double the size when did you start to bring in external help to help you with that recruitment process because obviously it can become very time consuming but equally you want to maintain that bar of talent and you don't want to lose touch with it too soon but you also want to make sure that your time is best used to, you know, add even more value to the business?

13:15Yeah. So we work with external recruiters almost from day one. And from my previous experience, recruiters are also a little bit tough. They are really good ones and there are a lot of not so good ones. So I knew a few really good ones for engineering and for salespeople or go to market people. I would say until we were 50, 60 people, we used external recruiting only. first one then we added another two three recruiters and now we have i think three internal full-time recruiting employees that handle that but we still use external recruiting too at the end it's a cost thing right recruiting is also expensive if you do it externally at one point if you hire a lot it makes sense to have internal recruiting also yeah awesome that's good to hear and that experience of knowing good partners in that instance recruitment but i'm sure there's been lots of other examples as you've taken people and partners and ways of doing things from previous experiences and applied that to Dash Zero.

14:13If we take a step back and we look at the software industry with the impact of AI, obviously with more code being written by AI and now obviously Dash Zero is supporting a lot with the observability, the role of an engineer is changing quite significantly. You're so close to this problem. I would love to hear what is your view of the future of engineering and how will teams look and what will they really be responsible for beyond just kind of monitoring and observing and directing the AI to do something that previously would take hours for them to kind of literally type out? First of all, to be really honest, I have no idea.

14:57I don't know. But I was watching in my team and we have really good engineers, how engineering or coding has changed over the past nine months. Since the new models came out in December and cloud code became amazing and then the others catch up, I think at Dash Zero, we almost do not code a single line of code manually anymore. So every line of code is generated by AI. And you do that by essentially prompting what you want to build into the system. So you are still coding, but in a different way with natural language. and then a lot of the details is handled by the AI and you become way faster, maybe 10 times faster in coding.

15:35But then there are other steps like reviewing the code, merging the code, putting the code into production. And those steps are still highly manual. There are some agentic approaches. So the work at the moment has shifted from producing code to more reviewing and quality checking, doing the quality gates. And to be honest, that's more boring work than creating. So that's one thing that's changing at the moment. But I think over time, this will also be handled by AI. So I can see that from prompt to production, everything will be automated. And then the role of the engineer is more understanding the requirements, understanding the architecture so that they can define the prompts, the skills, et cetera, what they need in a way that the output is good, right?

16:18Because when you do it yourself, I can just recommend everyone to just use Cloud Code. It's super simple, right? just put in a simple prompt and say, hey, I want to build a CRM system or something. And then you will see that it's amazing to get like to 80, 90%. But then the last 10%, if you get into the details, it's hard, right? Because then you have to start doing all the specific details into the prompts so that the eye gets you 100%. And I think that's the part where good engineers will become very important to gain speed. 100%. I mean, I've done a bit of vibe coding. I'm not technical, but I built little tools like you used one of them, PodPrime, to do my show notes.

16:58I built that on Lovable. But when I first started, I didn't even know how to construct a database. So I just built, not knowing really what I was doing, made loads of mistakes. And then the next project I did, I got it right. And you get better. But obviously, the real engineers, they're so much more knowledgeable about how products are actually built, how they should actually be structured, that even if they're getting the AI to write the code, they're writing the code for a product that is built in the right way that ultimately builds a better end product than a non-engineer who doesn't really know how these things actually work.

17:37Yeah, and that's why it's good you have done it. You have to learn it yourself to understand because you would think you just prompt and everything gets automatically, but it's not like that. No. It's still software development. You still develop the software. you're just not writing the code, which is a part that took a lot of time because you had to type and we were not super efficient in it. And there's also a lot of code that just is very easy to write. It's time consuming. I was a developer for like 20 years, so I know that. But then with AI, that gets beat up, right? And I love that because I never enjoyed doing the boring stuff.

18:10But you still have to know how to develop software. Exactly. Yeah. And there will be a lot of engineers listening who find all of that very interesting, but there will also be a lot of people who aren't. And my next question really is around how you as a business look at adopting new AI tools and how you manage it, because you have a large team now who are all very technical and savvy and aware of new products and things. But from a procurement standpoint, that must create its own challenges. So how do you choose which tools to use, which tools to approve, budgets per employee? How have you managed that kind of AI adoption yourselves internally without letting cost spiral and people ending up using different tools that don't work together and things like that?

18:58How have you handled that challenge? So at the moment, I decided because the world is changing so fast to not put any limits on my team. Nowhere. I'm encouraging everyone to use AI everywhere. We try to be AI first. So that really means, for example, we have the clear instruction that no presentation can be touched by human. So every presentation, every slide we have in sales and marketing internally must be 100 % AI generated. And that way you need to create that knowledge base and the style guides, et cetera, so that it works. And that's true for all my teams. And there's no limit on budget.

19:35I don't have to decide yes or no. They can just use any tool. Of course, we are tracking the cost. If it would be crazy, I would definitely look into it. But I think at the moment, people should just try and error. They need to test tools. We have no standardization. So a developer can use different coding tools. And we are switching around all the time because it's also changing, right? What was great a year ago or six months ago is maybe old school today. So I think that's harder for enterprises, right? Larger companies to adopt that quickly. But we don't have any procurement processes or people can just use it.

20:11And I think that's great that you're doing that for something like slides, which impacts multiple teams. But you know, if you're encouraged to do that, then the clawed skill or whatever you're using to generate them, if it doesn't create the perfect output, the work is actually to go and improve the skill that then translates across the entire organization, rather than sort of some people using it, some people not, some people going, why does this deck not look the same as that other one that we had? from you know that sort of thing so you by saying actually we don't want to hear such it you encourage the behavior to focus on building the underlying intellectual property almost into the workflow of everyone in the business it's kind of funny i was joking because what what happens if you become kind of ai native is that you're starting to document everything yeah it's kind of counterintuitive for a startup right we were always proud of that we didn't have to document in every process.

21:09But now if you want to have an agent do it, you have to basically write it down. Or as you said, you have to create a skill that generates the right presentation. And if something is wrong, you cannot change it manually. You have to correct the skill. But then the big benefit is once you have that skill in your repository, everyone is benefiting from that skill. It's super powerful, but it's something you could probably have done without AI, but nobody really did that so yeah i guess the only example might be companies sending email templates but obviously not every email is going to be a template whereas now almost everything should be almost like a clawed skill in some way that can be improved and tweaked absolutely yeah and what's the biggest budget you've had in a day is there a sort of interesting story there where you sort of wow wow, someone spent 50 grand yesterday or something.

22:03No, no, no. I mean, I know I look at the spend of my developers. So like our highest spend is around 20K per month per developer. I think that's already a lot. I mean, we are based in Europe, so it's probably more than the salary of the developer that we spend on AI. And then you already start thinking, what is the value? What do I get for it? Is there real ROI? That's why we also created a tool internally and we now released as a product called Darkplane where you can really analyze your token spend and map it to ROI because we needed it internally. I wanted to know, is that 20K spent correctly?

22:39If yes, if I get the ROI, why wouldn't I spend it? But I had no visibility into that. So not something, not like 50K a day or so, but 20K a month is already if you have 50 developers and everybody does that. Yeah. That's a million dollars, right? Yeah. I think that's really interesting though because you basically encouraged the behavior, but then built the tool that you needed to sort of check on it and check that it wasn't. Whereas obviously the bigger enterprises who are a bit more stuck in their way, they probably want the tool first before they encourage the behavior because they can't really get their head around the idea that one person might end up spending 20 ,000 pounds on AI credits.

23:19And it is probably also on a different scale, right? Think about you are, I mean, I'm talking to large US banks. they have 20 ,000 developers. If they spend$1 ,000 a month, that's 20 million. That's on a different scale, right? I can see why you want to have some checks in there because it's probably super scary if you give a tool where you could spend $100 ,000 a month easily and there's no control in there. And that tool, internal tool, so this is more of a question about how companies operate. Do you ever think you'll license that to anyone else Or is it just so personalized to what you guys need it for that you'll just keep it in-house?

23:59And is that the future of sort of running businesses? You build almost like artifacts or one-time use software or just software that's so personalized SaaS for your own business. Because I know Clio, they have an AI spend leaderboard. And actually, it's a goal to try and spend more sometimes. I know that PolyAI are tracking employees' AI spend as well. So everyone's kind of building these tools to solve their need. Do you do that frequently? Does someone on the team go, oh, we need to know what's going on. Let's build it ourselves. I mean, in this particular case, I would say yes overall. But in this particular case, we also released it at the product of Dash Zero, this DarkPlan product, because we are an observability company.

24:41And it made sense to have that insight into coding spend, because we can map that to the production data. If you think about ROI, you need to understand what's actually getting into production, how users are using it to do that mapping. And so we saw our opportunity for us to build a new product category here, and we released that a month ago. But other than that, we are building a ton of tools now also for special jobs that we will never release. And that's just an internal tool. And maybe we'll throw it away a month later. Yeah, the cost is so low of building it that I've seen even people building one-time demo pages.

25:16and it's like a whole functioning website but it's is it that too yeah it might get viewed once by one person and then never visited ever again but it's a fully fledged like website super interesting i think my takeaway there is if you build something for yourself but it aligns with your company mission and what you are known for selling to your customers then you could release it in public whereas obviously someone like clio they're a fintech personal wealth app they're not known for AI spend observability. So if they release their internal tool, it might not match up with the market and the brand and everything.

25:53So that was really interesting. Mirko, I'm going to move on to our two final questions. I know that you've done a bit of angel investing, but I'd love to hear of a future unicorn prediction. So a company that you've spotted that you think also has the potential to be a unicorn like Dash Zero. Yeah, and you said it must be a series A or earlier, right? Ideally, ideally. Yeah, so I would say Conduct, a London-based company. I'm also proud to be an angel investor. They just raised a$60 million Series A, and they built an AI operating system for enterprise software. So think about having an AI layer over your SAP system that helps you with changes, migrations, and asking questions.

26:36And they had a great start. It's a really, really exceptional team. JP, the founder and CEO. So all the ingredients are there to become a unicorn pretty quickly. Great. That's super interesting company. So thank you for sharing that. And then our final question is just a bit of fun to end the show. But if you could have dinner with any three people, who would they be? I was thinking hard about that. I would pick number one, Jurgen Klopp. You're British, you know him. But he's now a head coach of the national team also in soccer or football. And he's just, I think, an incredible, funny and good person.