SPEAKER_00 0:08

Today's guest is Estra Cajun, founder and CEO of HighRise, an AI company transforming how organizations understand and process unstructured data. A longtime AI entrepreneur and a former headhunter, Estra has spent more than a decade challenging the limitations of traditional recruiting systems. Today, High Rise processes millions of documents every month and is building AI-powered infrastructure that goes far beyond just simple keyword matching, helping enterprises turn information into intelligence. Welcome to the Scaling with Clarity podcast, Ezra.

SPEAKER_02 0:50

Hi, hi Maxime. How are you? Thanks for having me.

SPEAKER_00 0:54

It's my pleasure. It's a Saturday morning, hence why I'm looking so casual, but I'm looking looking for the conversation and uh and and better understanding what you're building. So why don't we start with the founder story? Kind of like how did you get into um venture and entrepreneurship?

SPEAKER_02 1:13

Uh that's a very, very old story. Um, so I graduated, I studied economics and I graduated in 2009, which the ones were a little bit older. Uh know that that was probably the worst time to be a you know uh an economics graduate. So I worked as a headhunter, um, and there I saw that we had multiple repetitive tasks such as scheduling, with um, and then my neighbor back then, he was uh a developer, so I went to him and I said, How can we automate this? Um, and we started together my first startup, um, and it was an AI-based or you know, very basic AI compared to what we have now, of course, but it was an AI-based scheduling tool, also an API like I have now. Um, and I worked on that for two years, and then I exited that in Holland to a CRM system. And after that, I took a break. But um, I worked as a uh partner at a venture capital firm for seven years where we invested in um you know, first marketplaces, fintech, you know, the usuals, but later on in our second fund, we went into deep tech, lab grown meat, semiconductors, lepgrow diamonds, defense force, very early on, defense force. Um, and um four years ago, I quit that job uh because I wanted to get back into entrepreneurship and I was looking into what we could do with AI um because I was following what an uh open AI was doing back then, before it was ChatGPT. Um, I was one of the early users of ChatGPT, you know, the beta ones, and I understood that a new wave wave was coming, and I wanted to tap into that. And that's how we started high-rise. It started as you know, extracting data from documents, uh, mainly for it's a platform-based HR businesses, and then we grew into finance and healthcare. Now we have seven verticals in total. So that is actually my entrepreneurship journey.

SPEAKER_00 3:30

Uh amazing. What um what propelled you to focus specifically on uh a word is essentially start building high-rise. What problem did you see that you're like, okay, no, this is worth solving?

SPEAKER_02 3:49

Well, I started high-rise because I thought I think a little bit like with the confidence you get when you're on the venture side or the investment side. Uh, I thought, well, you know, HR Tech is such a big market, um, but you know, we don't see these big, big unicorns in that vertical. So I'm gonna solve that. That was the problem that I was gonna try to solve. Um, and then we started off with you know, a workflow. Uh, the the initial goal was to build a SaaS based platform, uh, a SaaS platform. And early on, we decided to use uh a third-party parser. 70% of SaaS platforms or you know, platform-based businesses use a third-party parser. So we did the same. And we used the one that I used in my previous startup. And very early on, I saw that the technology hasn't changed, the accuracy was very low. So we needed, in order to develop something that is truly AI-based, as the you know, the standards of AI that we know today, so the extraction of data uh or the feeding of data into these models, um, it wasn't good enough. Um so we needed to build our own parser. Once we did that, I started to talk with other entrepreneurs, and they were like, Oh, well, this is actually really bug problems for us as well. If you solved it, can we use your parser too? And at that moment, we were like, Oh, this is actually a really big problem, and it's a true pain point because it sells itself. Um that's how we started.

SPEAKER_00 5:29

And so, what was the evolution from like uh okay? Well, we need to be the parser, the the parsers to read uh resumes, or what are we using the parser for? Because I mean, I I know that the in in the ventures that we've built, like we're trying to read through purchase orders, through invoices, to shipping documents, to builds of lading, through custom, and it's not that easy, especially because the information doesn't necessarily match. Like it's not like the if a product is listed this way, it's listed the same way across five different documents.

SPEAKER_02 6:02

Yeah, so we had, I think we build uh a database for close to two years, because you have all these variations on one word, so you need the whole family tree, the mother, the daughter, etc. Yeah, yeah, and that's why you start with you know just one vertical and then you grow into another one. And once you know how to fix every all the problems, you grow into another one. But what parsing is, and what we do, what we actually do is intelligent document processing, IDP. Um, because parsing is actually solved mainly. So the the extraction of one field that you find in a document, taking that data, word or number, that data, putting it in the right box, that's parsing. You know, it's sorting of words and meaning uh from a document, or you know, could be also an image, of course, but taking these words and meanings and put it in the right box, sorting them. What we do different is we connect a meaning to them and we also cross-check if that you know the information is relevant, if it makes sense, what the confidence score of that, you know, meaning that word or the you know box of text is. What that means is that, for example, we when you look at an invoice, you check if you know the subtotal is you know accurate, you know, all the line items if the sum total is correct, the amount time amount that needed to be paid in the time frame, and if it was paid, you can check that. And we can also label these documents so we can connect them later on, if especially when your SaaS platform is very important, because you want to connect multiple work uh documents within one workflow for a healthcare SaaS platform or a platform could be you know your in uh your invoice, your insurance policy, and then your lab results. Um, and then you have one uh idea. So I go to doctor, I get my lab results, insurance is paid, or it is covered, then I got an invoice. Um, so so that's what you need to do. So when you do that on a level like that, you start to develop something that is called in our industry intelligent document processing because it is intelligent, and then you also have um the capability to develop agentic AI or you know, highly more developed AI. So that is what we do. How we do that, we have a three-layered system. So the basic OCR parsing that you know is even open source available. That is also something that we use, but then we have the system that enriches data. We also have uh a confidence score, and then we check the accuracy of the total documents. We do that by token counts, um, and then we check if it's accurate. And if it's not accurate, we let you know that we were not able to parse or extract data from that field on a highly accurate uh percentage. Uh, so everything that is below 90%, we flag that we let you know that we weren't able to parse. That's also information for us because that's how we retrain our model. What are kind of like the I mean I mean we are an API-based startup, so you plug it in another, you know, technology company. It is very important to know where you want to visualize what we do, yeah just to let you know.

SPEAKER_00 9:43

No, no, I mean I mean I'm I'm loving this. We we're building a product at the moment, and this has been uh a big challenge for us. And and as I described, um what you see in one document is not it's not the same what you see in another document. It's it's the same product, but it's just the way that it shows up. And so like our ideal circumstance was to use a voice, uh, like uh a voice interface, whether it's 11 Labs or whether it's like uh WhatsApp, then have an intelligent document processing tool or something like that that kind of reads it and and in it can match it. Because our ideal circumstance is to be able to create purchase orders, update purchase orders, allow the teams to interact using their voice or text um to be able to do things. And um, we're seeing it firsthand, like you would request something and it was partially matched, and it would partially execute like well, I I couldn't match those things. And so um I think this would be well, at least in our case, this is a big problem because if we can create an 80% of purchase total, that's good, that's better. But ideally, I want to to get to 100% accuracy and get 100% accuracy every time. So, how would uh how would companies use it? Um, they would just plug it in within their own system as an API, kind of like walk me through what are the most common use cases, uh, or what are the first early first cases that that that use your API?

SPEAKER_02 11:19

Uh so the first use cases that we had were you know HR Tech startups, so CR uh no um applicant tracking systems and job boards used what we build. So job boards when a candidate applies for a job, they upload their resume, you need to populate the field, so you upload your resume. You don't want to you don't want to use lose a user at that stage. You want to keep them if I can't if I have to fill in that much, you know, yeah they they churn, they you lose them. So that's where they were used, and then later on, we were used in the matching of candidates because you know we extract data so well, you can also match that data them a lot better. Um, then we also have we started with uh Neuralink, and what they had was they had a lot of applicants. I mean, it's a top company, everybody wants to work there. I think they had millions of applicants, something like that. It's it's a couple years ago now. Um so they had you know, let's say a million applicants for all the jobs within a year. How can you match that against the 60 jobs that you have? That is, you know, humanly not possible. So they had with them, we build um, you know, a matching system that was based on the requirements that they have from people. With them, it's they want to see you know how exceptional you are. So they have the exceptional ability framework, I think. So they we build something that tests people's exceptional abilities by three really simple questions: look at the the you know, the data that we extracted from the resume, met that against the job, and then you know, eliminate all the people that or reject, you know, the people that were not capable of taking on such a heavy job because it's all really critical work that they do. Um, and then you know, and then do a one-on-one with all the people that you know had a high confidence score on the exceptional ability uh score. Um, so when we saw that, we started to see how you can you know take the data, take an you know, an AI-based model, so the intelligence of what you want to do, and then um implement that within a platform by using high-rise. Uh so today we have a lot more customers that you know are building an intelligent workflow within their platform, and to feed that intelligence, they use the high-rise, you know, document data extractor. The reason is because you get more data. Uh the reason is is you know, intelligence works as well uh the best when you have the the most uh information. Uh people who know, you know, have higher IQs, more information, read more books, watch more podcasts. Um, they are overall more in you know capable than someone who is not able to do all of these things. So that is the same for the intelligent models. Uh that's where companies now started to use high-rise a lot more. Um, and we see you know, agentic AI is fueling that even more.

SPEAKER_00 14:48

Agreed. Um, I mean, so many questions are coming to mind.

SPEAKER_02 14:52

Um, because maybe you have you have because the thing is that when you have more accurate data, you can let the agent do the work without you having to afraid that it's gonna fail because it's gonna, you know, say or think or uh think something wrong, make decisions, and then book some customer on the wrong spot or make a wrong decision. Yeah.

SPEAKER_00 15:19

So do you envision that uh every human is gonna have their own model that they've trained and perfected that the agents execute on top of?

SPEAKER_02 15:29

I mean, we already have that, right? Uh in what ways just kind of like the memory that's captured within uh within an LLM or yeah, I mean what you have in what you have in an LLM, you can especially with Cloud, it shouldn't be like an advertisement, but especially with Cloud, you can plug it into many uh other tools, you can have it answer your emails. I think if you work a bit on it, you can have it answer your phone calls, but there are tools to do that already. So if you cluster them all together, I have my own CRM that manages my life. Um, refrigerators become you know intelligent. I saw one, it you know, sees everything that you put in the refrigerator and then tells you what you can prepare and what you should throw out. Um, can you imagine something that tells me? And I could probably set it up now. Like you have to wake up at this time, you have to take these vitamins supplements to be healthy. Because last week you did your blood work, you have to achieve these goals for your career, you have to sell this much for your, you know, it it we all if you put it all together, we are already there. But the question is, is it intelligence? Not yet, it's not at that intelligence level yet.

SPEAKER_00 16:51

And so uh and I would agree with you, and and a lot of the information is orphaned. And can you put pull it all together? Of course you can. So, how do we get to that intelligence level where some something is making decisions uh for us on the simple things and just giving us a decision framework for the harder things?

SPEAKER_02 17:12

Well, the thing is that we we as an AI community that you know the people who work in this field who are trying to be groundbreaking and push the boundaries of what is what AI is capable of, and I think that we also have a role in that. Um, we have the responsibility of developing these models and making them more intelligent. So IQ in general. Um for that to happen, you need you know data centers, but you also need information to feed these models, and that's where we come in. So most of the models that we know, they've already consumed all the information that is available on the internet, yeah. But there is also a lot of information out there that is not digitalized yet, and I I I can I know that because we have big customers that you know digit that who are digitizing all the information that they have in boxes or you know, ancient Sumerian tablets. We have it all. Um, all that information needs to be digitized and fed into uh these models. I think some companies are gonna come with, you know, upload your we're gonna you can upload all your diaries and we're gonna make life story video. That's a hack, and then they're gonna take all that data and then train AI models like what the Pokemon Go uh game did. Um, all that data that we have, and I have a lot of notebooks, but all that data that is not digital, they're gonna go for that. And in order to extract that data, you're gonna need higher eyes, you're gonna need a tool that is able to extract that data with high accuracy. And you know, in a safe environment, we have a very safe environment, we never had a leak, we don't see information, we you know, we we have all the certifications. I think it's also very important to keep that data safe while you work with it.

SPEAKER_00 19:20

So I want to double-click on something what you said because I I couldn't agree more with what you're saying, and you gave the Pokemon Go example. And for those people that don't know, what Pokemon Go did is essentially you have Google Maps, and Google Maps get you to the building or get you to kind of the I live the environment in an outdoor setting. Pokemon Go allowed you to extend this beyond by you navigating around specific areas, and so it's an acquisition of data via uh it's essentially, I guess, augmented augmented reality. Do you see like more and more of this happening where we're acquiring data what it's through like AR classes or otherways, and and kind of using this to continuously enrich the data set that we have? And and yeah, go ahead.

SPEAKER_02 20:11

But do you mean by is you know, the data that we collect is someone is gonna hack, you know, what Pokemon Go did to you know cheat a little to get information from people? Well, it's a little bit cheating, yeah.

SPEAKER_00 20:29

Um I mean, what startup doesn't cheat in the beginning?

SPEAKER_02 20:34

Uh I think everybody cheats.

SPEAKER_00 20:37

Exactly.

SPEAKER_02 20:37

When I was an investor, I always said, what is your advantage that you know you couldn't tell just everybody, yeah, but you know, it's gonna get you ahead. What is your what is your heck? How are you hacking this? And it it's all in a legal framework, but there is a hack. So with Uber, I think it was, or was it Waze? Some one of those. I think we spoke with them and I asked them, not Uber, but another you know company that does the driver experience. He said, Well, we put these cars more cars on the map. So when you open it, looks like, oh, there are so many cars. And Uber also does that. You're like, Oh, I'm gonna be here in a couple minutes. I got so many cars, even when we know we see it and we're like, I even have that. I'm like, Oh, there's so many cars I'm gonna get.

SPEAKER_00 21:28

Yeah, yeah, of course.

SPEAKER_02 21:29

And I do get it in a minute, but it's not that car that I saw on the map. So everybody cheats a little bit within a label framework, it happens, and then you also have people who are you know afraid that their data is not out there and they want to clean out their data from the internet. I mean, you could do that, and I have great respect for everybody's choice, but I don't feel like there is much to hide for me. And if there's something that I don't want to be on that I don't want. To be on the internet, just don't put it online or on a computer and just write it down on my notebook.

SPEAKER_01 22:06

Yeah.

SPEAKER_02 22:07

Um, so it's gonna happen. Like there is not a lot of privacy. Uh just create your own privacy. But what if what what is gonna happen when people know my age and my you know my date of birth and place? Nothing. However, personally, I don't want my DNA sequence out there, so I don't do 23ME. Everybody has their own garbage. So there are gonna be companies who are gonna try to hack getting data, personal data or stories that is you know written in notebooks to get that out, get it online, and then sell that data. It's gonna happen. I mean, it's it's unavoidable.

SPEAKER_00 22:48

Yeah, but I feel like is it fair? No, well, sure, but but but but also I mean, sure, more and more data, uh, more of our personal data is available online, but also the advancement in terms of how data is protected and secured their advancing. Uh I mean, like how many apps do you use, like uh just passwords is very rare. You will have like a second factor of authentication, you might have a third factor, you may use biometrics, you might like. I feel like everything will advance in in a um commensured way where it becomes really hard for somebody else to use your data, or very easy, and then it doesn't become that valuable enough.

SPEAKER_02 23:37

If you know, if data is very easy to to access, and there is like a big flow of data off a customer profile like me, yeah, then I don't become that important anymore. But if it's the more we close it off, I think it's gonna be more valuable, and they're gonna find more creative or even sinister ways to uh find it. So I'm not against the data transparency, however, I would love for you know countries or companies to safeguard that data, but then again, I mean, we hear about countries that you know got hacked on a government level, and then the data of all the you know citizens was on the dark web. So there's not a lot of privacy there left anywhere. Maybe it's even better if there's just so much data around that it doesn't matter anymore because we keep the things that are very important on a different spot, different place, and we don't leak it out. But you know, it was I think a couple years ago, before this AI wave, I don't remember who said that, but someone you know, one of the top tier one investors, or we maybe Jeff Bezos, he said data is power before it was oil, and that is that is exactly it. At this point, it is very, very valuable. If you want to develop AI, you need to have data. Okay, and what we do is we help you find it. Well, we help you structure it, we don't find it for you, we structure it for you whenever you found it. We didn't have, and then coming back on the security point, if someone comes to us and says, Well, we have this questionable, very personal data about these people, and they don't launch it that that that way, but we understand that we don't parse that. That is, and that is at this point something that we don't do, and that's a decision of the CEO, which is me. But I think there are companies who will do it, and if you stole that data, then you're probably able to parse it at a certain level anyway. Yeah, so data is gonna be a big topic, and we're playing into that.

SPEAKER_00 25:48

Agreed. Uh so as we deal with with a ton of uh a ton of ventures, what we've noticed is that there's really only three um modes that that exist at the moment. How quickly can you move vis-a-vis your competition? What kind of access do you have to data, especially proprietary data? And the third one, um blanking out, so uh data velocity, um oh, distribution. Um, what kind of distribution do you have? And what we're seeing is companies that learn how to use these data, and I'll give you a real example. So major oil and gas companies. So think that the the super majors, what is like you said if you uh you worked in the Netherlands, so it's like what is Dutch or Dutchell, like a major, major one gas companies, and then you get the companies that were the service companies like Schlamberger that were essentially a service provider, but they were accumulating a ton of data. And there's actually a study by an Australian professor in Australia University that proved that the companies that accumulated data, even if this wasn't data, grew something like four times, five times faster than the company with the assets that were generating the data. And so to me, this is like there's no better example to to to show them to say, okay, well, if you wrangle your data, if you uh uh accumulate it, aggregate it, clean it up, like use it in a in an intentional way, there's a lot of value in this. I mean, and and you said, like, well, data is the new gold, and and in many ways it is. Um, are companies taking this journey or are they still like kind of like reacting to the situation?

SPEAKER_02 27:38

No, they they are. So, what we see is that I mean, this is part of our growth journey, of course. We need to understand what the market is doing, but uh what we see a lot is that in the beginning, especially when companies saw that you know they could replace work force with AI, and they're like, Oh, that's amazing. So the board members are like, Well, we can save money and become more profitable, and our stock value will increase because you know we're more efficient. Uh, what they most of them did was you know try and use an LLM because it's cheap and cheap compared to building your own and it's going to be easier. So they work in a demo environment. So when you have a small environment, it works. But as soon as you you know unleash it on a big data set or within a big corporation, such as just for as an example, it's not our comp uh our comp uh customer, but an SAP. It's a huge company, or an Oracle, or uh LVMH, you know, big companies that are globally known, multiple offices, big supply chains. Um, so as soon as you unleash it in an environment that is that big, then it just doesn't work. It becomes expensive, it becomes slow, the accuracy is low, and then you start to see that 80% isn't enough. And the bigger problem is it that it hallucinates. So when it doesn't have information, it makes things up, which is a big problem. You can't make things up in an environment that is that critical. So then they were like, Oh, maybe LMs are not gonna do what we do, and then a lot of these companies were okay, we have to build our own. So they used um, you know, for example, a reg model framework. Um, and then you see that that is also not enough because you still missed that part where you extract the data from all your all your data silos and your documents or you know, all the information. Um, so so that is actually something that where we add value. We be we are an API, so now it becomes valuable for us to slowly start to you know integrate into these um uh platforms that are that you know enterprises are building. So that market is coming, it's maturing, it's not there yet because you know, enterprise is a little bit well, a lot more slower because that's how they get disrupted.

SPEAKER_01 30:16

Yeah, yeah.

SPEAKER_02 30:16

Um and it's just the the the nature of that animal, it's you know, it's not because they're losers, it's just how it works. Um, but that market is coming, and then another market that is coming, and we didn't ask, but I'll just share. But is the voice and video uh part where you said I want to be able to say something to my CRM system, and then it does something for me, exactly. Yeah, and that has a couple years to go because the accuracy is just low. Um, and you can understand that when you watch these, you know, somebody does that, that's swiping. When you watch watch these, I don't watch TikTok, but I think TikTok also has that, and then Instagram and YouTube. You have the the the subtitle where where someone says something and then it says, I don't know, like oh, it's very cold, but it's just there is a slip of the tongue, and it's then you see it's bold, so it's not capable of you know translating that you know well. So you don't want to have these mistakes in enterprise environments, but it's coming, it's coming, of course.

SPEAKER_00 31:27

So I mean you you you bring in a really good point. Um, as I mentioned earlier, we we're using um either 11 laps or telegram or whatsapp as the as the voice interface to interact with our CRM systems. But the one thing since March, my interface has changed, it became voice rather than my fingers. And what I mean by this, I you I use an app called Whisperflow. Um, and what I love about it is like I'm in the top 0.1% of users, so like in what three months I've spoken 500,000 words, and so I find it extremely efficient to use your voice to interact with systems a lot better than using my fingers. And and and for me, like this is my personal observation. What makes the difference is that you provide higher degree of context because you speak at typically two times to three times uh greater speed than when you input it via keyboard, and so better context or more context leads to better outcomes. So, do you see this uh continue to proliferate and and and essentially change in your interface? What is voice, what is eyes, what is like. I mean, you said you work with Neuralink, kind of like what's coming next? Tell me.

SPEAKER_02 32:53

I mean, they're gonna put chips in people's brains and you know, have them have the computer in their head. It's not it's not a secret. Uh then they have the hardware to do that, you know, they have a lot of work to do, but it's gonna happen. It's it is there. They had two people, successful surgeries, and they're they're paralyzed, and they can now do things by even just thinking about thinking about stuff. So that's gonna happen. I mean, I remember when I don't know, I was 19 or something. I remember I had a phone call with a friend of mine, I had a Panasonic, and the thing that that phone could do or anokia, I don't remember, but it was just a normal phone, um, like normal phone for what we know. So two-colored screen, you know, the the the pixels and then the color of the screen, and then that's it. And then I think a friend of mine went to this tech conference and or somewhere because it information wasn't that freely flowing. Um, and he said, Yeah, you know, did you hear about you know what they're gonna do with phones? And I was like, Yeah, I heard about that. I don't even know where I got that information from. That's but I said, Yeah, you can see each other when you have a phone call, and it was like, Yeah, that's gonna be so cool. Yeah, and now it's like yeah duh.

SPEAKER_00 34:15

Yeah, yeah.

SPEAKER_02 34:16

So this is the same. So even I remember that phone call, and this is the same with you know, having your computer infrar interface implanted in your body, it's gonna be the next, yeah, it's a bit futuristic. And if you don't want it, you just don't do it. I don't, I'm not gonna do it. I'll just read more books and train, do more brain training. I just want to stay the full human. I like that more, but there are people who are very interested, yeah.

SPEAKER_00 34:45

But it but if you don't start adopting this, don't you feel that you're putting yourself at risk of becoming obsolete?

SPEAKER_02 34:56

I think that everybody's so extreme about everything. I mean, I have other capabilities that maybe the computer doesn't have, so I will keep leveraging them. Maybe I'm really good in gardening and growing crops, then I'll just do that. It's very nice to grow crops by the way. Um, it's very suiting. Um, but is am I gonna be obsolete? Yeah, maybe, but you know, we'll see. And I I kind of like being, you know, this is very futuristic, but I like being human. I do not prefer to go to Mars because I like this planet, it's the best planet in the universe, by the way. We're the only ones who have trees, no one else has that. Um, so it's I think it's rebellious to stay human.

SPEAKER_00 35:43

Yeah, Ezra, you've lived in many different places. What's your favorite place?

SPEAKER_02 35:49

This planet.

SPEAKER_00 35:51

Well, no, no, I I get this planet, but like if if if if you can be in one city um in uh in in the world, where would you be and why?

SPEAKER_02 36:06

I mean, it's I I can't do that. I can't be in one spot, you know, all year, nowhere else, it's just one place. I have no other place, or you know, no other group of friends that I need to visit because you know they have kids or whatever. I don't do that. It's it's nice to have you know a global way of living life, it's very empowering and it's freeing. Um, but I think Istanbul comes pretty close at the top. It's yeah, it's a great hub, everything is far, but close as well. Uh, it's very, very challenging, so you never get bored.

SPEAKER_01 36:46

Yeah.

SPEAKER_02 36:46

Um, and then you have the sea. Um, there is a lot, a lot going on. A lot of people travel through, like even friends uh from the US. I had a friend of mine, I was in Istanbul. He was flying from San Francisco to Sri Lanka, they just stop and they have a day in Istanbul. I think it's a great haul. I think it comes pretty close at number one, but not all the time, it's very tiring. I love Miami as well. I like Texas a lot because I like brisket, but you know, I love Holland because it's you know so yeah, real. Um, so yeah, it's it's impossible to choose. I love Italy because you know it's Italy.

SPEAKER_00 37:31

Yeah, um, can we just double-click on Istanbul? Um, I mean, there's clearly something in the area in Istanbul. Like, I I look at textiles opened three new cities uh as premier cities, and it was uh um uh in Nebraska, uh one of them. The other one was uh Sarajevo in Bosnia and Herzegovina, and the third one was Istanbul, and they that's where they see a lot of the growth and potential. Like, what's happening in Istanbul right now? Like, why is Textarist going there?

SPEAKER_02 38:08

Um, I think we could why would Textart go to Istanbul? I mean, I mean, there was probably there is an invitation, there is a sponsorship, there is money that they can make there that they do it because you know it's in the end of the day, they're an investment firm, so there needs to be a financial incentive. Yeah, after that incentive, and if that incentive is not big enough, then they wouldn't do so. Either the incentive is really big, and I'm like now very business minded, but either the incentive is big enough, or then you know it becomes there is an incentive and they believe the market. And when you look at you know, gaming companies in Turkey, you know, have multi multiple uh unicorns here, then you have a lot of marketplaces that were really big. So the consumer market here is in Istanbul is huge, yeah. Uh it's you know, 80 million people in Europe. Well, if you count it as Europe, it's in Europe. If you count it as the Middle East, it's still you know huge, not like in Egypt or anything, but it's closer to understand. There are more rules, there is more laws, there is infrastructure if you compare it to some of the larger uh Middle Eastern countries. Um, so so that must be the incentive, and then there is also a bridge to Central Asia that comes from you know uh Turkey. So I think that the exterior minister had a meeting with the exterior minister of Russia about you know the Central Asian trade development plan. If we put all the wars and everything aside, so there is a lot of going on with you know Central Asia, the bridge between that and Turkey. So it's it's it's a it's a hub, it's it can be anywhere, it can be anything. So I think that there is a good reason why Techstars then think, oh, we can find startups and entrepreneurs because it's a hub, it's a connection point, or there is a big incentive. So who knows?

SPEAKER_00 40:11

For sure, for sure. Um, no, there's definitely a lot of energy around uh Turkey and Istanbul. Uh, you have like famous investors like Bunish Pabried, like like for him, this is the number one place to look for investment because um the values are fantastic, the the the scalability is fantastic.

SPEAKER_02 40:31

So um especially consumer businesses. I wouldn't I don't want I don't want to get stoned, but personally, I wouldn't do B2B in Turkey, it's a very difficult because they do a little bit more different business than what we know in Europe and the US because it's it starts to become Asia and the Middle East, so it's more on personal relationships and you know, um but B2C is amazing here. Yeah, people have a more of an American, you know, consumer-based uh mindset. You could sell a lot more than you could sell to a German, yeah, yeah, yeah, yeah.

SPEAKER_00 41:12

Uh of course. Yeah. Um, I mean Turks are very well known for being a good, good negotiator. Um, and so it that's that commercial mindset, it's gonna run through the veins.

SPEAKER_02 41:29

Yeah, I think it because it's a difficult country to live in, and the city has you know, Istanbul is 20 million people, so it's a very tough city. You can't get around here if you're just a nice person. It just you can be a nice person, but you have to, you know, you need to ask for, you know, you need to fight a little bit more for your spot in line than with your average country. Um, and the people here are real hustlers, so they know how to start. There is this saying, I don't know if you know it, like start as a Turk, continue, finish as a German. They're really good at starting things off, they're really you know, passionate about things, so they're really good at starting things, and you know, startup is that mindset being energetic, believing, yeah, even though maybe you don't have it all, you know, but you believe in yourself and get there, so it's it's pretty close to what it takes to start a startup.

SPEAKER_00 42:25

Yeah, I've never heard of this name, but now remember it.

SPEAKER_02 42:30

Um, I think it's in Dutch, I'm not sure, but Dutch people say that.

SPEAKER_00 42:36

Start is the third, finishes the German, yeah.

SPEAKER_02 42:38

But it is like that. You see them always like very passionate and starting off really strong. Yeah, Germans are good finishers, they you know, absolutely and everything and well, it's a different.

SPEAKER_00 42:51

Um, what's um what was the biggest mistake you've made in uh in high-rise and kind of what's coming up next? What was the most challenging? What's coming up next?

SPEAKER_02 43:02

I think that most founders have the same problems. We you know, wrong hirings early on, um that costs money and time. Uh so having a team that works very well together, like a machine, is very important. And thinking like, well, this person is so good, maybe it doesn't fit the culture that we're trying to build, but it doesn't matter, or fits the culture, but it's not the best person. Yeah, uh you just make make a mistake because you just want to hire fast. That's that's I think the first mistake that comes up in my mind, and then the second mistake is thinking that you you don't, because there is always um you know a competitor or technology that is about to evolve. So every second, every day is very mission critical. Like Elon Musk has that urgency uh mindset, and he is right, you need To have that urgency because you know every minute costs money, it costs a lot of money. Um, so yeah, thinking that you have time is a big mistake that sometimes we all have that, we all we will keep doing that. But I think that that is one of the things that I made that was one of my bigger mistakes.

SPEAKER_00 44:22

Yeah, interesting. And uh what's coming up next? What are you building?

SPEAKER_02 44:29

Well, we're we we keep on building better document intelligence. What we're gonna do next, what we're gonna release this year is um one API endpoint. So you just send any document through one funnel, you know, you know, one endpoint, backend engineers would understand it better. Um, but one endpoint, and then you know, we understand which document it is, so we read it and we let you know, you know, in the JSON output, so the outputs that we get when we structure it. We tell you, like, oh, this is an invoice, this is you know a patient file. Uh so that is the first thing that we're gonna release. Next is the name change. We're gonna change the name high-rise. That's gonna be um, it's gonna be field because that's literally what we do. We we extract data from fields, um, so it's gonna be field, um, and then towards the end of the year, we're gonna you know roll out some beta use, or we're gonna roll out the voice and video parser for some of our beta customers or so it's gonna be a very busy day, and the team needs to grow, and then we need to close the seed round. So there is a lot of things to do.

SPEAKER_00 45:45

Undoubtedly, um, I know we're coming on time. I I like to uh close the interviews with a choice of a question. What's the best advice you have ever received, or what's the kindest thing anybody has done for you?

SPEAKER_02 46:00

Oh, um, I mean, like the kindest thing that anybody has ever done for me. I think that people I'm very lucky, people are very nice to me, privileged. I don't know why, but uh so I can name a lot of things. What is the nicest thing that anybody has done for me? I think I think my grandfather, you know, for believing in me, even when I was like, you know, 24 and thinking that you know I was the best CEO of the world, you know, big ego in your younger years. I think believing in me, what my grandfather did, that is the nicest thing because it really helped me, you know, having someone who trusts you is very important. Yeah, um, then you have like investors who put money in when you didn't have the numbers, customers who trust you, employees that you know you're like, Well, we can't pay this month because we haven't closed yet, but you know, keep on working. That is amazingly nice. I think my team is amazing, even when you're down, and when you're a founder, some days not a lot, but sometimes you get down, and your team is like, Well, it's you, get up. So there are many, many you know, niceties that I can and then what was it? The biggest advice that I have for people, the best advice, yeah.

SPEAKER_00 47:16

What's the best advice?

SPEAKER_02 47:18

I mean, uh the best advice is probably that I get is don't overspend your money for my dad's grandfather, but I want to give my own advice to people is that um if you set a goal, and then you know, you have that, you know, two advice you have that people say shoot for the stars, and then you get to the more well, it's actually like set a big goal that is really far, because if you set very realistic goals like I want to raise the seed round or the pre-seed round, your mind thinks, oh, pre-seed, and then I'm good. It starts to chill. But if you especially with the World Cup going on now, if you set that I want to get in the pool, you know, the first games, I want to join the World Cup, then that's what you do. But if you try to win the World Cup, that's a whole other journey. You prepare differently, you train differently, everything is different. You're like, well, one a game, you know, first three games. Doesn't matter, you have like 10 games to go. I don't know how many. Um, so set that big goal. Even if you stop halfway, your brain will keep working until you get at a really good place. At pre-seed, having a brain that is like, well, now we may now we can chill is not good. That's how you feel. That is the first, and the second one is that with especially because you do a lot on you know growth and GTM, um, is that I always a lot of people talk about, we spoke about that earlier. Um, is that people talk about the low-hanging fruits and you know, go to the people that you know you can talk to and stalk people on the internet, etc. Well, you can go to the low-hanging fruits if you know them, if you have them in your network. So make sure that you have a really big network. Be nice, be likable, be interested, and be interesting. Be yourself mainly, don't be crazy. Um, put yourself in situations where you can win, and that is very important. So make sure that you have low-hanging fruits when you need them. And it's very important to start like when you're very young, especially a lot more charismatic when you're young and then working hard. The older you get, it's yeah, like yeah, everybody's somewhere. So start when you especially start when you're young, be charismatic, go out there, check in with people, not a lot, but be there and be you know, make sure that people remember you.

SPEAKER_00 49:52

I couldn't agree more. I couldn't agree more, but it's a great um point to end. Um, one thing that uh somebody told me, I wish I remember who it was. Your network defines your networks. So how you build your network is is super important. Uh like I I recently um I don't read books, but I listen to books. Um, there's a book by Tina Sealing who used to uh um be a professor at Stanford, and she wrote a book on luck. I couldn't, I don't remember the exact name, but I can recommend it highly enough. This is a fantastic book about how to create your own luck.

SPEAKER_02 50:33

Yeah, I think Malcolm Glanwell has outliers, so maybe luck is a good pick for him to write on as all his last couple of books were that great. So with luck, yeah, like I also I really believe that I'm very lucky, and I am always very like my boyfriend. Is like, that's not gonna work. And I'm like, No, I'm very lucky, it's gonna be okay. I trust the universe. He's like, it's not gonna be like that. I'm like, see, and then it happens like easy gonna have it, yeah. Now he's always like, How are we that lucky? I'm like, because you know, I made you believe that you're also lucky. Like it, it it I think it's it sounds very corny, but when you believe it, that's how you look at the world. Look, I'm talking about things and I'm saying I'm lucky and everything is great and light, I'm fun and bubbly. Look, I have my goal points. I even had like points where I'm like, maybe God doesn't like me. Um maybe that's the reason that you know I'm in pain for like a year plus because entrepreneurship is very hard, yeah. Um, but you know, it it's very important to keep on finding that place, continue. And I think, for example, if you look at Elon Musk, he's like a child, like he's like, I'm sorry. Um, and I think that is something very important that helps him. I I think move forward because he's like so childish about things, also his failures. He can forget about it when you're more mature, you get bitter. The kid is like, Oh, I fell, but you know, here's some candy. I think that childlike mentality is very important. Lows are very low in entrepreneurship.

SPEAKER_00 52:16

Uh it can be high.

SPEAKER_02 52:19

No, not everybody wins. I lost, you lost, we all lost multiple times, and people remember only the good stories, and they tell only the good stories. So I'm like, Yeah, I'm so lucky. I'll tell you about good times that I wasn't. Um, but it's very important to keep the mindset.

SPEAKER_00 52:38

I I think people learn more from mistakes than they learn from success. Of course, of course.

SPEAKER_02 52:44

I mean, you also probably have that. Like, I have moments where I'm like, Well, this is tough. There is a lesson, I don't see it, and I hate that, but you know, let's continue. Like what Winston Churchill said. If you keep if you go through hell, keep on going. Yeah, you have to do that. Okay, it's tough. That I don't want to be this on my I don't want my last words to be this on the podcast, but you know, Winston Churchill is not that bad.

SPEAKER_00 53:12

Absolutely. Um, Ezra, it was an absolute pleasure having you on. Um, Ezra, the founder of Higher, soon to be founder of Field. Um, can't wait to see what you're building. Would love to continue the conversation. I think that um what you're building is is essential for companies. Okay, it doesn't matter what the industry you're you're in. Um, and so I think if you're very well positioned to be um at the forefront of anything that's coming up technology wise. So thank you.

SPEAKER_02 53:47

Thank you so much. It was great talking to you as well, Maxime.

SPEAKER_00 53:51

Absolute pleasure.

SPEAKER_02 53:52

Bye bye.