9:29
Hi. Hi everyone. My name is Eliana Magra. I'm a journalist with Kathimerini and it's a great pleasure to be here with you today. If we could have both our speakers on stage. Um So, we're going to speak with Victoria Dollie who's the co-founder and president at Vini and with Johannes Galazounas who's the co-founder and CEO of Diffraction about how we go from product to adoption. And um in preparation for this panel I was thinking, you know, what I would like to ask them that could apply to not just, you know, tech or startups or companies, but to life in general. And um I would like to start with that question which has to do with trust and mostly trust in yourself. So, I was wondering, as you both know cuz we talked on the phone, um
what it was the moment that you both felt that you trusted yourselves and your ideas enough to turn them into products. And was it harder for you to sort of sell that to yourselves or to investors afterwards? Um I can get started. Hey everyone. Thank you so much for having me here. Um this is my first conference in Athens, so very excited. Um I think there's two parts to this question. The first part is how do you build confidence in yourself to take a risk that most people would consider a pretty big one, which is taking a leap on your own. And then the second part is
once you do decide to take that risk, how do you decide on an idea? How do you validate that your idea has legs on? So, I'll answer both. I think on the personal side on building trust in yourself, it all starts, in my opinion, with small wins. And what I mean by that is putting yourself little by little out of your comfort zone up until you flex that muscle. So in my case what that meant was taking the leap and going to college in the states, going to Stanford, being far from my family. Once I did that, instead of majoring in economics, which is what I was comfortable and good at, I decided to do computer science with a
focus specifically in AI, not because it was particularly a great programmer or anything like that, but rather because I wanted to prove to myself that I could do it. And once you do that enough times, you build that muscle. It's It's really an iterative process that eventually you want to keep raising the bar and eventually you realize the highest in my opinion, the most challenging thing you could possibly do is build something from the ground up. I was always really drawn to that idea, but it only I only felt comfortable to take that leap after I had proven to myself step after step that when I set my mind on something, I can achieve it.
Yeah, that's fantastic. Um Yiannis here, Johannes, um co-founder of The Fraction. And yeah, also great to be back in Athens. Also for me, uh first conference in Athens, even though I grew up between here and Crete, um but have been away for quite a while. Um maybe on the on building also what you said, um I think that the first part of like convincing yourself you can start something like this uh came from exploring different things and not being afraid of exploring different things and taking these risks. I think it's a bit of an intrinsic thing if you are fine with taking crazy risks and doing changing countries. I've lived in five countries in my life, um worked across six seven industries in
completely different jobs, research, um but I had a bit of a common theme in AI, um and always build things from scratch for the last uh 20 years or so in the industry. So, um it was always kind of intrinsic to start a new thing. So, at some point, like you also took the decision to go to the US um to university to MIT um and in the middle of my career actually, so uh maybe not in the archetype of the super young founder uh that we talked about earlier, um but uh took that jump, uh worked in a looked into something new, which for me was um what is the upside of quantum on the sake of quantum technologies. I was in AI but not in quantum technology, so I went to uh university, studied uh quantum technologies with the people who
invented quantum technologies, uh which was pretty special, and then uh found out about uh the potential of these technologies, too, and spun out a company out of that. Um so, I think that's the first part. And then the second part is how do you believe then in that this particular thing is a good idea? And in my case, um I've been building products for various industries, but in AI, so I didn't always know a lot about that particular industry was building in um by nature of not being in that, but I knew about the technology part. So, um we built quantum cameras. I brought a little satellite here. Um I didn't work in space before, but I knew about AI and I knew about vision systems. So, the main point for me was uh you need to find a customer
who's convinced. So, go find like whatever I think about this doesn't really matter. The point is find the people that will actually care about this, and then um build something with them, show them what you have, and then keep building till basically you can actually sign first like a letter of intent or MOU or something, but you know, much more importantly that they're willing to invest in either in our case more like research grants um or some initial contracts that they really want to pursue. So, when you find enough people that validate that idea, then you know on the customer side, right? And on the technology side, it's a bit about um yeah, trying to understand what you can about the technology. For me, this is
coming together in such a so many technologies in this camera that I couldn't nobody actually can validate this on their own. So, we need to get a lot of data points. Uh and ultimately tinker enough till it actually works to some extent that you can project how it would work. Um so, if you get enough of those data points, then you're more convinced about uh the technology itself. Yeah. And the market and then the product market fit. Yeah. But technology is moving very fast and uh I suppose the way that we incorporate it in our lives is also changing in a very fast pace. So, how do you make sure when you create a product that
it's going to stay relevant um in the next few years or that the tech it's built on will not be antiquated in like a year or two? Yeah. I think that's where you really need to take a bet on the team. When we pitch Finny to our customers, which for context is a growth engine that helps financial advisors source and manage clients, we always say, "You're buying our team. You're really buying the fact that we are going to make sure that we stay at the front of AI. We work tirelessly. We are the intense founders that will always just stay, you know, like at the forefront." And that's really I think the winning argument
because the product that one sees today versus the product that they will see in a week or a month from now is might be totally different. Um I think the most important thing is really just to echo what Johanna just said, is to land on a problem that really truly matters to the customer. And Y Combinator has this incredible term for that. They call it a hair on fire problem. They call it a hair on fire problem because it's so pertinent, so pressing, so urgent that any solution will do for the customer. It doesn't have to be perfect. The idea is that anything is better than your hair being on fire. And ideally when you land on a hair
on fire problem like that, what you observe is instantly a willingness to pay. They're not, you know, trying to get this for free. They're rather, you know, putting their card down and saying, "How fast can you get me this product?" I remember for Finny, we when we were first getting started, we had we our first customer we told them, "You would be able to get access in May." And we said May because the reality was there was no product before that. And I remember them saying, "Well, if I pay you double, will I be able to get access in April?" And we're like, "Well, unfortunately not, but it was such a validating signal that we were onto something because the demand was just so clearly like proven." And I think another really good proxy
for demand is looking at what your customers are doing today in the absence of your solution. And in an ideal world, what you want to see is that they're hacking together processes. They're like spread between multiple tools that were not built for the solution. They're jumping from spreadsheets to email, to LinkedIn in our case. But before we go to Johannes, do you pivot? I mean, while customers are using the product, do you see how their demand is changing and then do you change it a bit as well? I think in the case of Finny, the problem statement was always the same. We always knew we needed to help advisors grow, and we knew this was the right problem because it was the first, second, third, fourth, and fifth thing
that our customers thought about when they woke up in the morning. But, the solution actually did change quite a bit. It started as this massive database where we aggregated 300 million profiles in the US, which is all publicly available, unlike Europe. We synthesized all the data together, things like your net worth, where you live, where you work, and we thought that product would help advisors find new business. But, in reality, that wasn't the case because not only did advisors need a database, they also needed a way to action on the data. So, then we ended up building this agent to automate the outbound. And now, since then, have expanded into more products. But, yeah,
the problem always stayed the same. Johannes, on uh the first question is what? Like, how do you make sure that Yeah, um so, for us, we built this first of a kind quantum cameras. It's a completely novel paradigm, really. And um I think it's explained a bit more from the angle of the deep tech. So, when you try to find a completely new category-changing technology, it's going to be early, obviously. If you're Otherwise, you're already late, right? Um but, if you're early, then you will be good in some things, like a niche that you have to find. And there has to be somebody like that who really, really cares um that because they don't have another way to solve that issue. So, um
you have to find them, and they will take uh you know, they will take the hit of usability, of you know, it doesn't work that the other thing that I'm used to, I have to change my workflows. But, with they don't have a workflow, and the problem hasn't been solved, you know, this is where you enter it. And um I mean, there's this it's it's quite common knowledge, the innovator's dilemma, it's like your know, you know, Silicon Valley book. If you haven't read it, you should read it. It's very interesting. Um but there's kind of a classic exploring that idea of um finding that niche. Now, for somebody to care about it, A, it has to be you have to solve a problem that nobody else has solved, right?
B, um because it's a longer bet, now besides that first problem, what else can you solve, right? Now, in this case, it's a camera, so everywhere you have cameras, right? There's potentially billions, trillions of devices that could use this. A bit louder. The microphone is off. Um and beyond that, um the question is how do you give um orders of magnitude improvement, right? So, can you make a hundred, a thousand X in something? And in this case, we did. So, we said, "Yes, that you could have that potential improvement if um you give it enough time to develop, right? And then you make it as usable as the conventional technology that exists today." Um so, once you find that intersection in deep tech, specifically,
I guess it's a little bit different, but in deep tech, you want to have this multiple orders of magnitude that cannot be reached conventionally. There's no other method to reach it conventionally, either. Um so, you're really betting on that potential um to bet down on it. Hm. Um let's stay with you for a second, and then we'll move back to Victoria. Um and I want to ask you about your failures. I mean, we're now talking about your successful products that have been adopted and are in the market, but did you have any ideas that did not materialize, that did not even become products, or did you Were you part of teams that created products that then the adoption was harder, trickier, it didn't work out? Um Johannes?
Um yeah, tons. Um because it's a camera, so we had to initially think about all the markets, um and I came I did some of drug research before actually I came into before I founded this company. So, my idea was microscopy. Turns out microscopy is not a great business to start because of all the regulation, you need a long time to get into and so on. Um and then we went to semiconductor manufacture with chips. And I went to a conference, talked about a hundred people or so. And one of them, you know, we went further in conversations and one of them said, "Okay, I can only process one chip per hour.
How many can you do?" So, we calculated it's about 20,000 per second. So, we said, "Okay, that's maybe too fast because you have to catapult it and it actually will land in orbit, the chip. So, it will immediately disintegrate and explode." So, we said that's probably not a bad not a good application. But what goes 20,000 kilometers per hour? Well, spaceships. All right, we should maybe talk people who are building spaceships. Um so, there was a I would say a few months of discovery of finding interesting but kind of useless applications for an initial beachhead.
Victoria? Yes. So, my original idea before Finni came from my experience at Uber, where I was on the product team and one of our biggest problems was around QA testing. So, we would ship a product that would be buggy, unreliable, etc. So, what I wanted initially to build was this QA agent that would constantly, you know, make sure that our product was not regressing in production. Um and the real difference between pitching that idea versus pitching Finni was really the customers response. In the former case, the answers were rather lukewarm. People would say vague things like, "Yeah, I could see this being valuable."
Not the signal that you want to get. Actually quite of an anti-signal. What you do want to hear is enthusiastic yeses. You want to hear that people have a burning need for the solution you're about to offer. Um and I think there's this framework that um is incredibly valuable when you're looking to validate your idea, which is called the Mom Test. And what it's saying is that people generally are polite and in being polite, they will often lie to you in order to not hurt your feelings. So, it's called the Mom Test because even your mom if asked, you know, about if presented a bad idea, will tell you, "Yes, honey, this is great." And the problem is that not only do people often lie to you, at the same time
um founders subconsciously, we often seek encouragement more than we seek the truth. So, More than you'll see, sorry? More than we would seek truth. Okay. So, we would ask questions, again, without realizing, we would say, "Would you use this? Would this be helpful?" Um when in reality, the questions we need to be asking to gauge real signal is things like, "How painful is this today? Why is it painful? What are you doing in the absence of this solution that I'm about to, you know, present to you?" And um it's it's actually rather difficult to gauge who's just answering in a way that pleases you versus truly feels the
pain. So, my first idea, I think, was very much in the camp of um [clears throat] yeah, not didn't have as much legs and also was something that had been tried a million times, which is 90% of the time a red flag. Um if you had to start distribution from zero again tomorrow, what would you change in the first 90 days? Let's start with you, Johannes. Yeah, that's a tough one, um cuz you don't know what you don't know, right? At the first 90 days. If I knew what I knew, obviously I would have done everything different.
No, but in hindsight, yeah. Um, I think we So, we started with um, government funding um, and we it was a research that was funded for um, 10 years or so by NASA and DARPA. So, which is the US uh, agency that does investments in early-stage defense uh, applications, but dual-use like GPS, the internet, right? These things were invented with DARPA sponsoring, similar to this. And we relied a lot on that funding, too. Um, initially we got that. And uh, we pursued just doubling down on getting all the government funding we could and get all the contracts. The problem is it takes a very long time. So, by the time you have even a commitment that you will get something till the time the money hits, it's a very long time.
Um, so I think initially we probably had reversed that part um, and basically seeked the funding first um, the private funding and then basically looked into parallel stream of government funding. Um, so we probably would have reversed that. They probably would have accelerated the first few months that we were, you know, discussing and dealing contractually with the government. Um, so I think in hindsight and probably good advice for anybody doing anything in that's government-funded, too. I think there's a lot going on in Europe, too, which a lot of government funding coming in. It's a double-edged sword, really. Um, one because there's a lot of
overhead for reporting, it takes a long time, there's restrictions to what you can do, what you cannot do, what you have to report. There's IP restrictions sometimes. Um, so yeah, you have to be really deliberate about it and you shouldn't chase the money, but you should chase that Does that contract help me achieve my vision of building a company versus uh, just trying to get the next round of government funding. Um, I think that's a key learning. Yeah. No, that's sound advice. Victoria? Um if I could go back in time, I think we would have just been more aggressive with our spending and During the first 90 days?
Oh, maybe not during the first 90 days. Um [clears throat] during the first 90 days, I think we would have we should have gone to even more conferences because conferences for us were incredibly helpful. So, In what sense? Because of the visibility or visibility and also meeting people in person, I think goes a long way. Like grabbing a coffee with someone, grabbing a meal, building that relationship, especially in our industry. It's a relationship business, right? Wealth management. So, I would have really doubled down on that and then I think coupled with that, I on the topic of not being overly conservative, I would have hired more engineers quicker once we realized that there was in fact demand. Um we had this wait list during Y Combinator of 200 firms who
wanted to use our product and we couldn't serve them because we were only, you know, three people at the whole company and only one was coding. So, not exactly a distribution problem in the traditional sense, but rather um distribution potential kind of going to waste almost because we weren't able to act on the demand as quickly as we should have. Which I mean sounds like a good problem to have, but I'm sure it was very stressful at the time. Um in writing we say that you have to kill your darlings, you know, to write a good piece. So, I was wondering if there were any features that you personally loved before you unveiled the product all or afterwards that you saw
that after it was adopted and after it was being used, you had to let go of. Um well, it maybe not the feature, but more of a use case. Um so, the first use case was very dear to my heart was microscopy um and specifically drug discovery. I think with putting a camera like this on a microscope instead of a telescope has the same effect, but um we'll be able to do some fundamental science um discoveries that would I think change a lot of how we do medicine in general. So, I'm very bullish on that. It was just not the right timing. Um and the second thing was it was So, when I first met my co-founder Saikat, he's a professor at MIT, he basically said, "I'm looking for aliens." I was like, "Oh, wow, cool. Sign me up. Um let's
find aliens." But um and that's a big telescope with NASA that uh is called the Habitable Worlds. We'll go in the 2030s out there. And it's a very interesting uh application, um but obviously that's still a bit long-term. NASA had some funding troubles. Um now going back again with Artemis and the new and ISAC, which is great. Um but uh there was maybe another darling we had to postpone or two darlings we had to postpone, maybe not completely kill, but realistically the most exciting applications we could have done initially have to wait a little bit. Yeah. Oh, thank you.
I think in our case, um having started in early 2024, we overly relied on AI. We kind of really thought AI was at a better place than it was. We essentially tried to make our product fully agentic, and the models just weren't there. The hallucinations were 100x more frequent than they are now. Um we really had to roll back a lot of the functionality that was meant to be autonomous and introduce the human in the loop in a more tight way. In a way, we were kind of too optimistic, too ambitious. We overshot it, and then we had to revert back into an interface that was a bit more UI forward. Example, instead of letting our users query our database in natural language and in plain English, we reverted back to having a drop-down menu
and selecting filters, which at the time felt like you know, someone had like stabbed me in the heart because I was so excited to use, you know, the latest and greatest. Um and now, funny enough, only, you know, two years later, we finally have moved back into having a more agentic product that requires less and less input from the user, but early days, um it was frankly a mistake. Um and as we're nearing the end of our discussion, I just wanted to get more practical and help our audience um understand, you know, you both talked about problem-solving as being the key to creating a product that lasts through
the test of time, even if time is relevant because it might be for a short window, um but also so that it will be successful with users and popular. So, is there a specific thought process that leads one to an idea that really provides a solution to a problem? Is it something that you feel both of you now could do in other sectors, too? Or is it just key to know your sector very thoroughly or one sector and then identify where is the gap? I mean, how can people in this room who might want to become entrepreneurs acquire this thought process so that they can identify the gaps that might lead to successful products?
Um I would agree with a point that Andreas made earlier, which is that experience in a field is by no means necessary unless you're obviously inventing deep tech. Um in fact, I would say it could, you know, hold you back because it prevents you from questioning everything and seeing everything from a fresh set of eyes. What I would look for in a market is really two things. Again, in the pursuit of trying to find a problem that people deeply care about, I would look for a market where there's certain macroeconomic conditions that are helpful. So, in the case of the wealth management space, in the US, you have this very powerful movement where so many advisors are now leaving the banks and are going independent. They're leaving the Goldman
Sachs's of the world and they're setting up their own shop. And when they do that, the first thing they need is obviously growth tools to attract more clients. At the same time, you have this immense wealth transfer, it's called the great wealth transfer, where 83 trillion dollars are changing hands and are getting passed down to the younger generations. So, a ton nine out of which, by the way, of heirs want different a different financial advisor than the one their family has been using. So, this money in motion opportunity was unique to this market. That's on the macro side. On the technical side, I would definitely look for an idea that previously wasn't possible. Something that only now technology got to a point that
you know, it's it's possible. Would you use AI for brainstorming? Totally. 100%. I think AI is the best co-pilot of all times. I wish I had AI when I was ideating, for sure. Um And do we think, and this is a small parenthesis that it can help lead to original ideas if it's just trained on things that have already been possible. I think with enough context and pushback, the one thing that obviously we all know is that it tends to agree with you quite a lot. So, I would push it to disagree with me, but I would use it more as a market research tool rather than an idea generator. So, I
would try to understand, you know, market trends and also the capabilities of AI cuz right now the best ideas will most likely be in the AI category. So, just realizing what AI is capable of and what therefore AI can do that previously wasn't possible. Sure. Thank you, Johanna. It's my also my COO for the longest time was between Claude and GPT. You know, they had slightly different personalities that I programmed to be one more questionable, one more supportive, depending what I need in the day. But, um the idea was I think, yeah, I fully agree that's extremely useful. There is maybe plug-in MIT had this disciplined entrepreneurship. Um there's actually a little chatbot that walks you through these 25 steps or so of entrepreneurship
that you do initially to find a good market. So, actually once you tell them a few ideas, it just spits you out the entire structure. So, if you're actually looking for a startup, it's I think it's free to use. You can find it. Um then find but, you know, the B chat is still whatever AI can predict something, but, you know, you have to go and obviously find a niche and find customers who care in that niche enough. And often that's not the niche that you thought. In our case, we had seven niches or so we were exploring and then we figured out for multiple reasons that space was the most um interesting too for it.
Um so, that's for the market. For the technology, fully agree with the cross-section of technologies. I think um the true treasures are in the cross-sectional and cross-discipline, uh a subject. Obviously, I chose hardware. We are at the intersection of we have AI, you have quantum, you have space, you have optics, you have photonics. There's a lot going on. And no single expert could understand that and say, "Oh yeah, that's a good idea." But if you bring four, five people together, then from different disciplines, you can actually pull this off. And that's what we did. Actually, that was the main reason I went to MIT because MIT is open doors. You can take any class you want. And I found those four people
that were willing to build this company together. And you know, that's a special place. I think if there's something to learn also maybe for European ecosystem and beyond, I did not encounter this. I my some education in Oxford, in Germany, in Greece. This intersectionality doesn't exist that strongly. I would encourage, you know, find these intersections and push these intersections. Thank you so much. And I think this is a great note to end on. Um, both in terms of the need for teamwork and I mean amongst people, but also amongst sectors. I think if we all talk to each other more and collaborate more, great
ideas and great companies can be built. Thank you very much for having me here with us. Thank you.
