David Villalon, CO-Founder & CEO @ Maisa

0:00David Villalon:First guest is David. Welcome, David. Thank you. I'm very happy to be here. It is an absolute pleasure to have you. You've just announced an absolutely amazing fundraise. But before we get into it, can you just give a quick introduction to yourself and what you're building? Yeah, for sure. I'm David, the CEO and co-founder at Maitre. What we are doing is quite simple. We are enabling companies to onboard and deploy digital workers capable of doing end-to-end knowledge job tasks, but real complex ones, the ones that need to read the file, check into systems, do cancellations, and come back with the answers.

0:36So we are the Enhanagentic process automation platform and addressing this completely new undiscovered market.

0:43David Villalon:Amazing. And I see in a lot of the comms, especially around the fundraiser, you talked about how 95 % of kind of like enterprise AI sort of like adoption projects have failed. Can you talk a bit more about why you think that's the case and why you guys will succeed there? So we will succeed, I think, because we've seen, we've always addressed AI very differently. The main reason of why AI fails is because it's by definition unreliable. That means that when you use AI, it doesn't mean if they are just tele-lamps or agents, which are built on top of tele-lamps. They come with intrinsic limitations.

1:20So the chain of thought that they make is unfaithful, so you cannot trust it, so you cannot verify the results. They have hallucinations that you cannot control. And the main problem is that everyone is using AI to get responses, so you are always opening the door to these problems to happen. In the other way, we, since the first day, decided that that wasn't the way to go. And that if you use AI to get these responses, you will always get these issues. so you will jump into a new problem, which is you can automate document extraction, but you will always have to put a human to review it, and it takes more time to review than even doing the task because you cannot trust it.

1:55So we decided to go another way. And you say not to get the response, but to get the execution trace, which means the steps that need to get executed to get to the response. So we call that KPU, Knowledge Processing Unit, and it behaves like a computer. So instead of using it to get this answer, we use it to tell us, okay, this is the first step that you need to do, and to compute and to execute to get to the answer, and this goes gradually. So at the end of the process, you get with this auditable result that you don't have to trust it or not. It's absolutely what has been executed to get to the answer, and you don't leverage the AI to get the responses.

2:34And thanks to the environment that it's running, it's by definition hallucination resistant, because if it makes up the code, it's not going to compile and it's not going to be executed. So we are able also to overcome those challenges. So it's literally a very fundamental perspective to address AI.

2:52David Villalon:Yes, it's almost like taking a step before you get to the response, breaking down the process step by step and making sure those process steps are absolutely correct before you then even thinking about the response. Is that right? It's more, it's a good approach. It's like what is happening today as a metaphor example is if when you were into the school and you wanted to do a math exam, if you come up just with the answer, which is what models are making, you will have the, and you don't go with the rationale that brings you to the answer, you wouldn't pass that exam. And that's what is happening today.

3:25In our case, it's different. We are computing the answer. So we are using the AI to tell us, okay, the next step you need to do is this one, execute it, and it's executed. Okay, now the next step is this one, and it's this one, and it's this one. And by doing that, we don't use the AI to get the response, but to guide us how to compute the response in a way. And by doing that, we have full control of what is happening. We can audit the results and we can also deliver that to the user so they can comprehend and trust what is happening under the hood.

3:56David Villalon:And can you give some use cases of maybe the industries that you're targeting or the types of process that you're helping companies or planning on helping companies kind of automate? Yeah, yeah, obviously. So who will be benefited to having auditable results? First of all, the regulated industries. So that's why we have addressed that market first. We work with some of the largest banks worldwide. So we work in banking, we work in insurance, we work in energy and manufacturing, mostly those industries today. And the use cases that we are doing are mostly, for example, trade finance in banking.

4:34we are doing also full tax reconciliation by the end of the month or by the end of the quarter and of the year we are doing use cases that sound simple like power of attorney extraction which is extracting information from a power of attorney document but that is not simple because if you just say yeah for that you cannot trust the results and then the same time gets but because we are able to verify the results it's also getting into production which is what we are seeing that's for example on the banking side then for the infrastructure side we have review of contracts renewals and management of infrastructure stocks supply chain monitoring and management so those are some of the use cases that we are already addressing with clients today and working in production

5:19David Villalon:and how easy is it for you to kind of deploy your AI your tech across different industries Are you having to focus on those core industries that you mentioned, or is there an opportunity to roll out to lots of different industries? There's an option to roll out to different industries, but we purposely are focusing on those because we have already identified very big return of investment use cases, and we are just scaling those. But our approach to the market in that end is that if you wanted to do automations in the past, It doesn't matter if it's with past RPA or with new tools that are today in the market.

5:57You have to be technical. You have to define these workflows with boxes and go the things in the steps. And maybe now you put AI agents in the middle also. And the problem is that the people that do these tasks that I just have told you are non-technical. They are people that really don't know how to do those things. So because we give them this platform, they are able to become citizen developers. They are able to use our platform with natural language on board these workers, which open us a lot of opportunities to land and expand inside of organizations. So we land into trade finance and we expand to wealth management and other parts of the company.

6:36Or we land into supply chain management in supply chain and then we expand to the HR team and to the finance team and we expand through the companies. but currently we are addressing the ones that i just have told you which is banking energy manufacturing mostly because we are finding very strong product market fit

6:55David Villalon:amazing so just two questions on that who are you selling to in these organizations like who have you got to convince to kind of get you guys on board and then secondly who are you actually who's using the product i mean it sounds like anybody within those departments is it like an operational team is it it could it be somebody in customer success could it be in i don't know Anybody using the processes? So those two questions. Who have you got to convince? And then who's using it? At the end, for those corporations, you don't just have to convince one person. Because they are large corporations.

7:28So you have to convince a lot of people inside. But usually you have to convince on the top-down perspective, you have to convince the people, the CFO, the operations people, the ones that understand really how the business operates and works and see where are the flows. so they can see how they can leverage and overcome the problems that they have today. But you have also to convince bottom-up. You have to make the people be excited about using the tool and showing that they can do it. Because if not, you jump into the problem of RPA, that it's that they cannot do it, so you have to hire externals to do it.

8:02And the way to address that inside of teams, you have to focus into operational teams mostly, so you don't go to sell really to AI and data teams. You go to sell mostly to the ones that are managing these processes and this automation and already have even experience with RPA and automations. And the way that we scale internally is, as you said, we usually find a few champions. They are the people that get it really fast, that are able to showcase that this doesn't just work, but also returns huge value. And then gradually you start expanding through the organization. We also collaborate with external channel partners, which are consultancy firms that help us also to hyperscale in these very, very big companies.

8:41But the focus is that any citizen developer is able to or any person can really onboard a worker to help them to automate the task that they are doing. And it's still a long way to go. It's a bigger problem that it seems to be done. And something that I can tell you is that once you solve the execution, what comes next is even more valuable, which is everything related to governance, performance evaluation, human in the loop for the use cases that did not work as expected and you need to have a validation or an evaluation of what has happened and all of that, I think that is going to be also a big thing for 2026 that we have already kicked off because it's what comes next.

9:26David Villalon:Yeah, it makes sense. And I want to just go back quickly to the fundraise, right? Because you announced that last week you raised a pretty mammoth $25 million seed round led by Creandum, but you had participation from ForgePoint Capital and a few others, NFX. Not that long after your seed round, right? Your seed round was, pre-seed, sorry, was the end of last year. I think it was$5 million in December. And now eight months later, you've raised$25 million. What were the key milestones or unlocks which kind of enabled you to raise a round of this size?

10:04at the end it was two things it was we had happy customers that were open to not just speak with the fans but to speak well about us and they were very happy with the results that it was one of our focus and secondly that we saw an absolute increase in demand that we needed more capacity to address we saw that we were getting more inbounds just because of the word of mouth between the industry and that we needed a way to hyperscale this because we found that it was really working. And there was also, in the sales process, I could tell you, or in the process of more than new clients, we found new playbooks that are not done that worked because we leverage our unique advantage of being extremely fast to adopt, extremely fast to develop.

10:56So we are very aggressive with that. We are like, all the ones are telling you three months. We are on Monday, and Wednesday you have the use case solved. And things like that. We are very aggressive, and this is what enables us also to create this, we can make it, and to create this momentum. And I think that this has helped a lot also to raise this, because we think that we know things that the market doesn't know yet, on a few sites, and we want to double down on those, and keep speeding up.

11:29David Villalon:Amazing. I mean, that sounds like both you're moving at insane seed, but also that's an amazing kind of problem to have, right? Being stretched for capacity, word of mouth, referrals. You know, you said you've got a lot of momentum. You're being pulled a lot. Are you being pulled in a certain direction or to a certain geography? I know that you're headquartered both in Spain and in the US. Are you being pulled more to the US? Is that where you're seeing most of your demand come from? Yeah, we are seeing it more today in Europe also because we push for it. In the U.S., we do have some operations, but there is a big difference.

12:02In the U.S., you have a market that has a lot of noise. Everyone is telling the same things. Everyone is trying to fight for things like this, and it's very challenging to navigate that market because of the full noise that is. While if you are in Europe, you have serious customers that really understand the businesses and that want to do it, and we decided that it was a better option to learn here while deploying value to our clients while we prepare for hyperscale into the U.S. So right now we are more focused into Europe, Spain in this case also because obviously here we are locals and it's way better for us to compete, obviously, than also in Europe.

12:41And we are right now, while we are speaking, we are now bringing more people into the U.S. to start seeding the expansion that we want to do there. But the difference is that instead of going there to do discovery, we will go there with a mature product, solid understanding, working playbook that will help us also to hyperscale. And this is also why we raise this round, to expand on Europe and start opening the market of the US.

13:08David Villalon:Amazing. So you're just getting into the US market. It's great to hear that you've had that success in Europe already. I mean, a lot of the founders that I speak to, especially in AI, they say that one of the problems across Europe is procurement. and procurement is both risk-averse and very time-consuming, especially when you're dealing with enterprises, that they have a very long sales cycle and they're quite apprehensive to take on risk and sign new techie young startups. It sounds like that hasn't been your experience. Is that right? Yeah, not really. It hasn't been our experience in that sense.

13:41David Villalon:No, that's amazing to hear. And I want to talk a bit more about those plans. I think I read somewhere that you're looking to double your headcount and use that money to hire. What kind of people are you going to be hiring? Are you looking for more technical talent? Is it GTM? Is it based in Spain, spread across Europe, US? Can you touch a bit about those expansion plans? Absolutely. So, yeah, we want to, as I told you, we want to increase capacity. And to increase capacity, we need more. On the product side, we need more musical inside of engineering side and also what is called product development.

14:15So we need more people working in the end. On the AI side, so we are talking about research and apply AI. We're a deep tech, not really common also in Spain and also in Europe. But we need more people there because we want to leverage the advantages that we have. So that's on the engineering and technical side. On the product, we are finding out that UX is, I think, one of the missing pieces of the market overall, of everyone. I think that we have still a lot of work to do there. So we are including more resources in trying to understand the UX, which is not the SGI, as you probably know. It's way more than that.

14:55It's understanding the full journey, the communication, the language. And obviously, if you want to grow its capacity, so it's growing on the go-to-market side. We are talking about AI automation engineers, forward deployment engineers, SDRs, people that will help us to grow also on the marketing side. And the good thing, also for everyone that is listening, it's that we are not focused just in Spain. We indeed have a lot of people working already from the US, but also from Europe. So we do hire remotely. We think that right now the talent, it's remote, and we welcome remote talent and people that want to join us around the globe.

15:39David Villalon:Amazing. That's great to hear. Can I also just ask, you mentioned that you're building deep tech. This is not like a lot of the AI startups that we read. This goes deeper than that. It's more closer to infrastructure. What's it like building a deep tech at the moment? Are you seeing the same type of hype that there are for other AI companies? And specifically, what's it like in Spain? No. In Spain, there are very, very, very, very few companies doing this. In Europe also. And what it's like is that when we were racing the first round, it was very hard to do it in Europe because it was hard for us even to explain at the deep tech.

16:20And it was hard to find. I'm talking to you on the first stages. You innovate and it's very hard to explain why this is powerful or not. Now it's easy to see it because of the value that we can bring. It's challenging because also we have very contrarian perspectives of a lot of things. Just to put you one example. In 2023, at the end, before MISA was founded, my co-founder and I were already saying that RAC, for triple augmented generation on AI, wasn't going to work. And while billions were bet on that direction, I believe you billions, and we were saying that's not going to work. So it was, I think when you are working really, really deep and trying to innovate in the, it's also, it's more challenging because then you need to translate that into value for the customers.

17:10So it's challenging, but also if you land a good innovation, obviously creates an absolutely, it opens a window of opportunity that you can take advantage on. So that's why I really like to be on the core as a research company. And that's why my co-founder is a scientist and we are a research company.

17:30David Villalon:Amazing. Yeah, you're really building your own tech. And yeah, it's great to hear that you have conviction and uncommon or unpopular beliefs. Yeah, I think it's easy to sort of jump on the bandwagon and try and build something that everybody thinks everybody else wants. But final question, you know, what is the top priority for you over the next six months? Is it GTM? Is it servicing the demand that you already have? Is it hiring? What is going to fill your time for the next six months? It's improving the, keeping improving the product. I think that we are not even halfway there. We still have to do a lot of things.

18:06It's improving the product and learning on the UX. It's, as you said, delivering to our clients today, to keep delivering to our clients, making them happy and learning from them, which is what we are doing, and obviously scaling our client-based. So we have a lot of people waiting to use us, a lot of conversations open, so it's been able also to increase the amount of clients that we have. So that's mostly product, making our clients happy and happier, so they keep speaking well about us and opening to more clients and scaling our operations. And then the go-to-market side. things like this one, like having people to know about what we are doing and that there is in Europe one new player which is not that new but that it's coming very strong and that there is a new wave coming

18:58David Villalon:Another big name in European AI tech and I'm all here for us David, thank you so much for joining best of luck with what you're building, I'm going to be keeping an eye out and if there's anything that you want to announce anything you want to share and you want to come back on, just let me know but otherwise, yeah, I'll be rooting for you Thank you, much appreciated thank you for your time My pleasure, thank you too, bye bye