Transcript: Jonathan Gur-Zeev | Weavy.ai | How to Build a “Design Machine” for Organizations in the AI Era

Eyal David: [00:00] Hi, this is Eyal, welcome to Product Builder, the home of the most interesting product conversations. I'm the owner of UserFlowzz and, among other things, the host of this podcast, and my goal here is for us to improve our product skills together. If this is your first time here, I hope you enjoy it, and if you've already listened and gotten value, I'd really appreciate it if you shared the podcast with a friend. So let's get started.

What Weavy is and why AI isn't in production yet

Jonathan Gur-Zeev: [00:28] Hi Jonathan, how's it going? Good, fun to be here. Very fun, Porto, nice to meet you. We're recording here from a, so to speak, complete studio. You're basically coming to us from Weavy, as CPO and one of the four founders. I'd also like you to give a short explanation of Weavy. Basically, great. So Weavy is basically a platform for image and video making, content creation, for marketing, gaming, basically anyone who produces content professionally. We don't target consumers or hobbyists. We target professionals, and it basically lets them combine any existing AI model, and any that will exist in the future, with traditional editing tools, and wrap that whole thing into what we call a design machine. Or we tell you: instead of designing a frame or designing a shot or designing something, build the machine that knows how to design it for you, of course in combination with AI. Okay, can you tell us a bit, so what problem did Weavy come to solve? Yes, so we all know that over the last two or three years a new model comes out every day. Two or two and a half years ago DALL-E came out and everyone's jaw dropped. Truly amazing things, and this happens on a daily basis, every day we hear that a new model came out and we're all very impressed by it. And maybe the most significant observation that came up for us, each of us partners from the direction he came from, is that in fact in the truly professional world of image making and video making, the whole AI topic stayed at the level of, I'll say, almost a curiosity. It did enter work processes, but more in the worlds of inspiration, moodboards, getting directions. And for production, for real marketing materials, let's say in the gaming world and in the brand and entertainment world, AI's penetration is minor, and that's exactly where we would have wanted, or at least the promise was, that AI would cause a very significant disruption. So two and a half, three years into this, and basically it looks like nothing has changed, and certainly at the business level you can't say there's a significant ROI, or a very significant change in ROI, when they use AI in the creative world. So that's where we started, the direction, maybe the insights we had about why it's not really, sorry, production grade. Part of it came from my background and that of Itai, one of the partners. In my past I came from the world of post-production, and Itai continued there, that's how we met, and he continued there for 20 years as a creative director. So the first thing is that it's very non-deterministic, meaning you don't know what you'll get from the AI, and that's also the magic of it, but you don't have the control and the ability to get consistent results every time. So that's maybe the first and most significant problem in working with AI in professional worlds, and maybe what separates a professional from a non-professional. Of course there are many differences, but I look at it in the context of whether I have a client I'm working for, whether I'm a freelancer or I work at a company and have a manager, versus hobbyists, where whatever comes out and looks tasteful to them is fine from their point of view. So maybe that's the first distinction. So professionals require control and the ability to edit, or precision, in their outputs. That's one thing. The second thing is that if you look at this AI ecosystem, then of course in the consumer, hobbyist worlds there are endless tools and a new model comes out every day, it's literally like that, every day some new capability comes out, and I basically don't have... I mean, I can work with five or ten tabs open and move from model to model, but tomorrow morning there will already be a new product and I'll need to learn how to bring it into my workflow.

Eyal David: [04:56] Let me ask you something for a second. So basically, when you started examining the opportunity for a minute, let's call it that, okay, you're saying there's also the matter that I want my result to be high-quality enough, and also to repeat itself at the same level of quality, and also basically that the process be workable, so that I, if I'm now the user, in this case the professional as we called him, won't have to work with lots of open tabs and a lot of guessing to reach the final result. And what basically distinguished between the hobbyist audiences and probably the ones who will pay is basically that there's a client at the end who needs to receive... he doesn't care what you use in the middle, he needs to get the result like today. Exactly. Okay, so you basically identified the potential here, and you also identified the huge boost that's really happening on the technological side of it.

From Fiverr to Porto: how the founding team came together

Jonathan Gur-Zeev: [05:44] Where did this find you in life, where were you? So luckily, let's call it, or unfortunately enough, I took a period for myself of so-and-so months to understand what I want to do in life. Before that, maybe we'll talk about it later or now, I managed a product group at Fiverr and I decided I needed a moment. I knew I was going to found another startup, I didn't really know in which domains and what it would be. And I call it, I was searching slash waiting for the opportunity, for something to come along that really fits my skill set and the things that interest me. So really, let's put a pin in that and move on, like, in brackets that we've now closed. So you were in some period, and then you basically joined the other founders? So Itai, the partner, one of the partners, I've known for 20 years. We're very, very good friends, so we're in touch anyway, and Itai also came from a startup he founded two years earlier, and the two other partners are friends from Fiverr. Oh wow, okay. Yes, we're a trio of ex-Fiverr people plus Itai. What are they, more on the technical side basically? So, yes, so there's Lior Albek, who's the CEO, Yonatan Alomut is the CTO. I'm responsible for product, and Itai is mainly responsible for marketing content. He's basically Weavy's user number one, really the ideal persona from our point of view. And they're all, they're all in Israel? They're all in Israel, all in Israel except me, yes. And basically, so okay, so you founded it, I assume that somewhere after that you also basically moved from Israel, right? Like, who knows. No, no, I founded it here, I mean

Eyal David: [07:39] Weavy started when I was here, yes, in Porto. Okay, and at that point in time you already had a view of the market, that there's an opportunity here.

The problem: designers use Midjourney only for inspiration

Jonathan Gur-Zeev: [07:48] Did you see it, did you start interviewing people, did you basically rely on your own knowledge, how was it? So we really started from feelings, from personal familiarity with these pains. I, again, it's been many years since I worked in this, but I do come from these worlds, and Itai does deal with it day to day, and the gut says, like, this doesn't make sense. Like, it can't be that this amazing thing, this amazing technology, we're not managing to bring it to production. When I say production, I mean the content people produce, images, video, audio, that really reaches deliverables in the end. Yes, exactly. And we didn't know from the start, was it clear how to define the problem well? Yes, I can say it today, also not in two words, but it came more from the gut. And yes, we started talking with designers and with videographers and understanding why you aren't using it. So everyone tells us, yes, of course I use Midjourney. What do you do with it? I make myself all kinds of references, moodboards, styles I like, and afterwards I do everything I do the way I'm used to doing it for 20... or it's not the specific person, but it's probably a traditional industry of 50-60 years. And we tried to understand, okay, so what's missing? Why don't you take Midjourney and put it into production? Ah, because it's not exactly what I need, or not exactly what the client asked of me. I have ideas and stuff, but it's kind of not, it's not there. Okay, so what do you do? If I could take it and refine what it does and iterate on it and keep working on it.

Eyal David: [09:36] Kind of like raw materials. So for those who don't know post-production, it's the people who basically do the process after the shoot of a production.

Consolidating tools, and a "design machine" instead of manual design

Jonathan Gur-Zeev: [09:45] I would do post-production and yes, use it. And from there came the idea. Now, also learning to work with AI isn't that trivial, okay, people usually know prompt engineering, let's call it, but there are many more tools that AI gives you to actually control things. But you need to know them, you need to learn how to work with them. And it brings me back for a moment to the point of tons and tons and tons of tools, each one does it differently, meaning every AI product like that solves maybe some specific problem. But it's a different interface, the terminology is still in its infancy and everyone calls things differently, and the ability, or the need, to consolidate was very, very, very clear to us. And I'll say the third thing is that, okay, say Photoshop has Firefly, yes, right, there are AI tools there, so you do what you do much faster. But the question, and it's a bit deeper, that we ask ourselves is: is that the promise of AI? That I'll do what I do just a bit faster, or twice as fast? So we look at this and say no. If I'm already building with AI, and if I'm already using the technology, why shouldn't I do it a thousand or a million times more than what I'm capable of doing today? And that's how this idea was born of basically, instead of designing or instead of editing, build a machine that designs and edits for you. Don't compromise your professional integrity, don't settle for whatever the AI spits out. You say what it needs to be, and once you've created this machine, now you'll get endless unique, unique outputs, but always according to guidelines that you, as a designer or video maker, chose.

Eyal David: [11:40] Okay, so let's explain this machine, okay? So basically I, as an organization or as the professional, define a set of guidelines I won't compromise on, it's basically the brand and the rules of the brand, and basically within that, go generate for me all the assets you can generate that basically meet these guidelines. And you call it a machine because basically I can now write, I want assets for ad x, y, z, from such and such an angle, in such and such sizes, and all this goodness will be waiting for me at the end and I'll be able to send it to the marketing people. Just saying, that's the rosy dream.

Jonathan Gur-Zeev: [12:15] Yes, so what you describe is a bit simplistic. I think I call it a machine because it can run again and again and again, and I'll get non-identical results, but it's not just a brand style guide or brand guidelines. It's me, as a designer, how I chose this shot to look or how this image will look: composition, color palette, details, depth of field, whatever, everything that people who make images and video care about.

Eyal David: [12:42] Over that he has full control. Really my design line, absolutely. So I still keep my say too, so I become like a professional times 10, for example, because of the efficiency in it. Excellent. I assume that your background at Fiverr, all the partners except Itai, if I remember correctly,

Jonathan Gur-Zeev: [13:01] so basically it also helped in understanding the problem. So, so say Lior, basically I managed the sellers' product group and Lior managed the sellers' organization on the business and operational level, and we worked closely together for quite a while. So Lior brought it from another angle. Lior, before he worked with me on the sellers' group, managed Fiverr's Logo Maker and got to meet lots and lots and lots of designers. And he really saw that, like, AI is not... okay, people sell "I'll do something with AI", yes. Of course it doesn't matter to anyone that it's done with AI. Which, by the way, is what Fiverr is now announcing everywhere, right, which is fine, it became a buzzword and, like, everyone wants AI, so why shouldn't I do it with AI. But zero impact on the sellers' business. And again, why? Why is there no impact? So it comes back, it pretty much came back from his side too to the problems we described: lack of consistency and lack of control, endless tools, endless new technologies to learn, all of them, and so on and so on.

Eyal David: [14:17] So, so we saw it also from the side of Fiverr's sellers. That's also another thing, you can think about this audience. Again, there are probably many audiences here and many industries, let's call them. Because it can be a video editor, and it can be a photographer altogether, and there can really be lots of aspects, but they're all already loaded with lots of tools anyway. Now adding all the other tools on top is something they don't want. They're learning lots of things anyway, and AI is a big thing that suddenly fell on them and they're also afraid of it, so there's also this mental aspect. Which you're basically also solving along the way, if I understand correctly. Once you consolidate all the different things into one product, and you're probably also doing some kind of onboarding or

AI FOMO, a "Zapier of AI" and AI literacy

Jonathan Gur-Zeev: [15:01] a very good and simple walkthrough, then you're also taking care of that along the way, right? So I'll say, we call it AI FOMO. Suddenly there's some new word you don't know, like, and what does it do. And enough, in Weavy you understand, ah, okay, so they took this model and connected it with that capability, and I can do it in two clicks in Weavy, and I too have the ability to use, let's call it, a new technology. I think it's a bit reminiscent of Zapier, what they do, because they basically took the software world, or the world of the various APIs, and okay, at the click of a button I can connect things. You're basically going for something like that, at the end of the day, raising... It's kind of, right, like very, very, very roughly it's a Zapier of AI. But what's more important to us, or I'll even say it, the company at the technological level deals much more with the conventional manual editing tools, adapted to the new world of AI. So if part of it is, let's call it, Zapier, then some of the components there will be AI, and quite a few of the components there aren't AI at all: editing tools, tools to create correct compositions and the right effects on each layer, and so on and so on. So yes, it's Zapier, but it's not only for AI. There are two challenges here, and we learned this too while working with customers and users. So at the education level, there's one level where people simply don't know what can be done with AI. They just know there's Midjourney, like, they know how to write something and something comes out, but they don't know the rest of the tools and the rest of AI's capabilities that exist. That's one thing we as a company need to deal with, we call it AI illiteracy: I don't know how to use AI, I don't know what tools even exist. And then there's the second challenge of, okay, now how do I do it in Weavy. So not all those people you described earlier... they need to learn to use, to know what can be done with AI, and then they need to know how to do it in Weavy. Those are two challenges that we as a company also need to know how to solve.

Eyal David: [17:16] Here too it sounds like, okay, as long as I focus on the end result of what I want to achieve, then the product can probably already guide the user and say, okay, this is the way to do it, like you're used to. There are also options here. It sounds to me, it sounds like a really interesting problem and domain. And before we continue for a minute on how you maybe did the exploration and how you moved forward, I'm really curious, when you came in, what did the landscape in the industry basically look like, what did it look like?

Jonathan Gur-Zeev: [17:45] So it really depends on how you look at it, okay, let's call it that. So there's one view that says, let's look at the existing image and video editing tools. So of course in these worlds we have, and I'll split this for a second into the professional world and the, let's call it, I don't want to say hobbyist world, because it's still, let's say, here, and for example these aren't people... they use it for work, but they aren't professional design people, okay.

Eyal David: [18:14] Or their clients, maybe by the clients you can distinguish between them, right? Probably yes, yes, so you can make that distinction.

One use case, or a platform for every use case

Jonathan Gur-Zeev: [18:22] And then in these worlds, obviously Adobe is the biggest player, and there are also the worlds of Autodesk and quite a few others. Maybe another view is to ask for a moment in terms of ROI, or what's the real value that Weavy provides, and if so, then I'm already entering the world of marketing automation, you could call it, although we're not necessarily in marketing, which is also an interesting point to talk about. Let's open brackets. Yes, let's open brackets. So, so I think this is a dilemma we had from the start of the journey, which says, like, let's take one use case and solve it really, really well: products in e-commerce, whatever, video of... the episode with Cloudinary, so a shoe, a shoe in 3D in the scene I want, and let's solve the best thing for e-commerce, for people who sell products. Let's solve splash screens and mobile games. There are endless use cases where we could have gone and said, let's do this the best. I can't say today what it was, like, why we chose what we chose, but I think it's more that we felt like doing something a bit bigger than solving a specific use case, and let's build a platform that basically knows how to do all these use cases. If the user wants, and knows how to use the software to build his use case, he'll build it, and let's not spoon-feed everything into the... I'm curious for a second, and what did the founder who comes with the marketing hat think about this? So, so there we had Lior, he's the CEO. It's not that he came with "let's do this", but he raised this point: why don't we come and solve a specific use case. Maybe it's worth saying that before that we said, okay, let's not make a decision now. Let's bring in users, let's see how they use it, let's see what they do with it, and little by little we'll learn better where we are, where the fit is more right. Not exactly product-market fit, maybe product-use-case fit. And we discovered that people use it in all kinds of ways for all kinds of use cases. And we said, let's leave this decision a bit longer, we'll postpone this decision. And I think that today we're maybe no longer at the question of whether we want to solve a specific use case. Maybe the question is more at the go-to-market level, what's the, let's call it, the industry, and from the industry derive the more correct target audience. But again, that's at the discovery level for us.

Eyal David: [21:10] That sounds, from the outside, yes, like a process somewhat different from the methodologies we know. I hear, I hear that, A, you know the problem space very well from your past, that you basically make many, many assumptions along the way, which also sounds great to me, and when there's no traction that's what you do.

Skipping mockups: design partners and real users

Jonathan Gur-Zeev: [21:36] Are you still working that way, or was that something from the start of the journey and... No, so today we have design partners we work with. And we have lots of private users, not private, who are freelancers, and we work with them to hear what they came to do, what use case they came to solve, whether it works for them, doesn't work for them, what they're missing, of course, in how the product gets built, right. So, so already, yes, there are still assumptions we make, but, but we validate them much faster than in a situation where you, like, without... you don't have a product, yes, you start talking maybe in user interviews. We skipped the stage of producing mockups and showing them, and I personally don't really believe in it when you're doing zero to one, like, at the startup level. Maybe inside a large company like Fiverr it works much better, because people have context for what you're talking to them about, and here I'm, like, coming from nowhere, and the only way to test it, from my point of view, was to develop something that isn't even an MVP, maybe we can also talk about MVP, but let's start seeing users use it, okay. Like actually running a campaign to pour users in? No, it's really, really people we know. Yes, I strongly believe that first impressions are very, very important, and therefore it's very important to control who enters the platform, who sees it for the first time, how his first-time experience will be, and we learned a lot about that too.

Eyal David: [23:12] Let's tell about that for a moment. Okay, you have lots of friends from the industry, I understand. So okay, you've now brought in Moshe, let's say, and you also picked him with tweezers because he works with such-and-such an enterprise and he understands the problem space, and now you got on a Zoom with him from Portugal and showed him, okay, this is the context, this is what we're trying to build, and then this is what we built.

Jonathan Gur-Zeev: [23:33] Yes, totally. So, so I'll say that even today, when we bring people in, we do one-on-one onboarding for them. The onboarding is a bit more structured, we already know what to say and where, where the pitfalls are that new users can experience. But we did, yes, we'd bring them on one by one, give an explanation, let them in, talk with them again. At the beginning it was very hard because everything was very shaky, like, in terms of the platform. I'll say we also watch user recordings, super, super, super helpful, yes. But that, that was the way to learn what works, what doesn't work, on what, where we need to invest. I'll say that luckily we weren't very far off at the basic level of the premise we set out with. Yes, lots of refinements at the experience level, and that's still our way to learn from everything. I really love this, because it's a get-real approach: we basically built something very fast and shaky. We as Israelis always really love it, after all it also itches to build

Eyal David: [24:37] and not go on a whole exploration journey, and then let it see real eyes as fast as possible and get that feedback. What challenges me a bit in thinking about this is that you, I, you bring friends, and then, like, it could be biased. Like, you need to manage this discussion really well so as not to get things that are friendly and nice. But if you know how to manage this process, and you're probably also doing it at a larger scale, then you probably actually get good results. We skip all the stages of, hey, I don't have a product yet in the middle and I'm only showing mockups, and that whole stage, which I also don't really like. I'm forced to do it many times and I don't like it. And okay, so you did this stage and you already have something built, so now you basically started to launch it. Right. And basically what was the next thing you did?

The LinkedIn post that brought in NVIDIA

Jonathan Gur-Zeev: [25:27] So lots and lots of things happened around it. Yes, luckily for us, let's call it that, we asked users not to post publicly, because we're aware that the competition around us is probably... we talked about the big ones, yes, we didn't talk about the small ones, and I assume there are a few more teams working on things similar to what we're doing. So we asked to keep things kind of discreet, and then someone posted a screenshot on LinkedIn. Wow, how fun, I got access to Weavy, blah blah blah blah blah blah. From that post, NVIDIA came to us from the United States, a studio with 200 people.

Eyal David: [26:06] And then we also learned what they need. How lucky that she did that. I have another question before we continue, the marketing aspect, is that something you thought of maybe adopting as an approach?

Jonathan Gur-Zeev: [26:18] Of telling people not to post? The opposite. Come on, post. No, of course. So okay, so we can talk a bit about go-to-market, so maybe put in a bit of the right context. So we're building software, meaning an app, yes, it's a website. I call it software, we're building design software. Content creation software for professionals, and therefore it has to be amazing. In its experience, in its look, in the value it provides. So that's one point. The second point is that if designers don't love it, then it won't help that I convince NVIDIA's CEO to buy 200 Weavy licenses. It just won't happen, and so obviously it has to come from the bottom up. That's for sure, and it's good that you're going with this approach, yes, especially since you skipped all kinds of stages along the way, which is great.

Eyal David: [27:14] But suddenly getting this feedback from bigger players significantly accelerates things, if I go by your approach of, like, let's let it see the real world as fast as possible.

Jonathan Gur-Zeev: [27:26] So that's the thing to do. Absolutely, you need to control the narrative. You need to know when you agree that someone, when I agree that a manager of NVIDIA's studio will see Weavy or try it out, okay, because he sees a demo, and in the demo we know what doesn't work and so we avoid it. When do you bring him into an experience where you know he'll succeed on his first visit to Weavy. So tell me for a second how it was, I'm really curious how it really was, that okay, you got this email back from NVIDIA. We started seeing, we have a kind of waitlist on the site, our site is very sparse today, by choice, and suddenly there started to be, like, someone from NVIDIA signed up. Okay, cool, let's send him an email. We sent an email, yes, we saw the post, we'd be happy, like, to do... if you'll let us do a demo for you. So we did a demo for something like ten people, a week later another five people, a week later another ten people, and suddenly we started getting waitlist sign-ups from all kinds of places at NVIDIA, and then we also somehow connected with NVIDIA Israel,

Eyal David: [28:38] and little by little it built up. Excellent. And now if we go back for a second beyond this story of her sharing it, so you started to see sparks, that people are using it, that people are signing up, and was it around a specific industry, did you already start distilling that? So still not, it's very hard for us to put a finger on, like, who benefits from it most,

Jonathan Gur-Zeev: [28:59] whether it's actually the gaming world, or maybe freelancers, and maybe in-house studios. By the way, most companies today have an in-house studio, I don't know if everyone knows that, but, like, what used to be exclusively agencies, today a lot, a lot happens in-house, and the work is similar to agencies. So, like, in the end a designer is a designer, whether he's a gaming designer, true, it's different, true, the needs are different, but his work tools are identical, they'll all work with the same tools. So still not, I don't want to get there, I want to be the platform for everything, yes. I believe that little by little we'll learn to sharpen our message better for certain audiences and certain use cases, but it will always be at the level of, I hope it will always be at the level of marketing messages and not at the level of the product itself.

Eyal David: [29:54] And today, if we look at the product itself, then as a user who comes in and gets access, what do you basically steer me to do?

Jonathan Gur-Zeev: [30:02] Okay, a really, really good question. We're exactly at this stage. I think the biggest gap in the product right now, at the level where we can open it up further to more and more users, is the topic of onboarding. Right these days we're working on onboarding. Until today, as I said, you book a meeting, a one-on-one session of 25 minutes with one of us to go through training. But once there's onboarding, which is also a process that will never end, again, I said earlier about education: it's learning the world of AI and it's learning how to use the world of AI in Weavy, so there are lots and lots of challenges in this world. It's not onboarding, maybe let's call it a whole education track that we'll constantly need to work on. Which will probably be very personal. And it'll be very personal. So I think that, again, if you're with people who come a bit from a background of, yes, knowing software a bit, it's not Photoshop, okay, but it's 3D worlds and compositing worlds, and people who come from post-production usually very, very quickly understand what they need to do and how to use it. Yes, there's a challenge, because again, it's not a website, it's not an app, it's really, really software, and there are many tools you need to learn, and which tools suit which use case or the task I came to do. You can also choose... different people will do the same thing with different tools. So, so there's a learning process here, or a learning curve, that's probably a bit higher, steeper than what you're used to when you enter an app.

Eyal David: [31:41] Look, but also when you were doing, I think also now when you're doing these one-on-one conversations, you probably start with, okay, what are you coming for, what's the goal now, the task you have on the table, right? So I think it'll be the same way, right? Probably yes, yes, interesting. And when you heard these jobs to be done from the people you interviewed who were basically starting,

Jonathan Gur-Zeev: [32:02] what basically shone above everything, what was the thing you did? Again, Weavy isn't built for the designer who's looking for how to get something done, how to take as much work off himself as possible, it's not that. I mean, the people who love Weavy are those who love to work, who love to be very, very hands-on about what they do. They don't compromise on the output, and then usually they know what they want to do, meaning they have a task of, now there's an ad they need to make, or it's some new screen for a game and things like that. In the marketing world of course it's much simpler, like, now there's a banner to make or now there's some social post to make, so it's very clear to them what they need to do. If I condense where we need to bring you as a user when you come to do something, usually it'll end with some generation you did with some editing, so that you know you can also edit what you're doing, and it'll be built in stages, meaning it'll be built at different expertise levels, where we need to keep taking you along. Like, it's not once you entered Weavy and that's it, but you'll get more and more and more information from us in a way that lets you really learn everything that happens there.

Eyal David: [33:30] Yes, gorgeous. It just really interested me. You know, in the end you need to make a really good first impression, so you have to, like, take the person along, or at least complete one task, so that's probably a real challenge, it's interesting. Tell me, and basically afterwards I also run my business inside Weavy, and then that's basically the hook, the thing that will basically keep me in. I think it's intriguing to know also whether their desire, this AI FOMO as you said, like, needs to be satisfied, maybe that's also a hook you can ride on.

Jonathan Gur-Zeev: [34:04] Right, so, so you raised a really, really good point about the outputs and the assets created in Weavy. So of course, the more such workflows a company builds, and the more design machines and more assets it has, the more significant the hook. But I'll say something we haven't gotten to yet, okay. I think maybe it's a problem where people haven't yet understood that they have this problem. Today, in this AI era, when you can produce endless content, and not that expensive, and soon it'll be even cheaper and so on, companies and teams and individuals will each have at minimum their own Shutterstock that they generated and created, tagged, cataloged and so on. And the question is how you leverage this crazy amount of content in order to use it. Yes, like, to me it always feels a bit like, it feels a bit anti, like, recycling, because you very easily make more, but why do you need to make more? I go back to the point, I think there's a real problem here of: I have endless content, what do I do with it? And we want to solve that too, and I think that once we solve it with a real solution, it becomes a very significant hook too, that all the... think Google Photos, like, where do I take all this stuff?

Eyal David: [35:40] And I think that's what's going to happen to every company, on steroids, like crazy. Yes, separating the wheat from the chaff and moving on. Very interesting, that's for example something I didn't think about at all, and it's a problem you probably didn't know about at the start when you set out, and it came up, and surely there are more like it. There are more things like that where you basically, as you described, skipped mockups. I just said that because, by the way, there are more things where you're, so to speak, kicking the bucket against the familiar methodologies.

Intuition vs. methodology

Jonathan Gur-Zeev: [36:11] So let's talk a bit about methodologies and about the word and what it means. I think a good product manager or a good entrepreneur, by definition, uses his intuition. Now, his intuition doesn't mean he feels like doing something or that it seems to him like the way. Intuition is built from many years of experience, of learning, of the desire to learn, of the ability to learn, familiarity with the customers, familiarity with the industry, familiarity with what you do, familiarity with the technology, and a good eye for UX, what works and what doesn't. That's intuition, okay, it's your experience, it's all, all the tools you've collected over the years. And if you have good intuition, you don't need a methodology. I'll say something a bit, like, not nice: I think a methodology is something written in hindsight about a very, very successful use case, leaving aside all the use cases, all the times that didn't succeed. And even the successful use case was probably not only related to the methodology the product team worked with, but to lots of other things, with some percentage of also luck. Yes, talent, skills, marketing, this, but also luck. And so I don't think a methodology as a stand-alone can hold up. It can make processes more orderly, and in some way, when a company is big, I won't belittle the fact that a company is big and there are lots of people and everyone basically needs to be aligned on the same thing, so yes, the more you do routines and work in an orderly, methodical way, probably the better the chances of success. Is a product manager who knows the methodology better than a product manager who doesn't know the methodology? In my eyes, absolutely not. I think there are many other things I'd be happy for my product manager to know how to do, and methodology isn't one of them.

Eyal David: [38:16] So basically you rely a lot on scars and on experience. I also connect with that a lot. Is there any other specific methodology that you also skipped?

Jonathan Gur-Zeev: [38:24] I think, you know, there's, say, user interviews, okay. So I think there's a right way and a less right way, and a methodology of how to interview correctly, and there are books about this. And I think that basically the fact that we could interview people in front of the minimal product that existed gave us the ability to look for a moment not at whether it's cool or not cool, yes, to hear what they say, like, on the periphery of things, or in their facial expressions and what they see. I think the most interesting measure is: okay, we brought you in and heard your feedback, but now that I'm, now that you're not talking to me, I want to see how you behave. And then the miracle of recordings and so on. So, so again, there were no mockups along the way, there were no user tests along the way, the user tests happen, happen on the product, like. And I think that... again, I don't call it luck, but I think we built a product from the start in such a way, with the intuition and the abilities and the skills and the talent of our team, that allowed us to reach a point where you can get real information. Not at scale of course, there isn't much data,

Eyal David: [39:42] but real information, where you don't need the methodology to explain to you what the right thing is. Yes, I think also, really, as you mentioned, at a small company size it totally makes sense and accelerates processes. I connect with it a lot in many ways. I want us to flip the record for a second and you tell us a bit about Jonathan's challenges. What are your challenges today, apart from this cold in Porto?

Jonathan Gur-Zeev: [40:12] Yes, no, I actually wouldn't define that as a challenge relative to what's happening in Israel. My challenges, so I think that today, today I do product work and also programming work together with other engineers, meaning I'm doing it, adding working hands into this thing. What do you have, actual tasks you complete, or is it more on the side of what you called prompt engineering? No, no, no, I, I'm really building the thing, building the thing, yes. In fact, I started it alone and slowly others joined. I started the code completely alone, I learned to program while writing the code and designing and defining the product. So I think our significant challenges today, and maybe it's also something we learned along the way, okay, the challenge, which of course also affects my challenges, is what to deal with now, correct prioritization, okay, because we do want to bring in more customers, we do need to close with them. I'll open brackets to say that the topic of AI, copyright, legal and all kinds of things like that turns out to be a hard problem there too, and we need to know how to solve that too. So, so we move between several tracks, constantly juggling between these things: between building the product toward the launch, supporting the design partners and seeing that together we manage to keep improving the product, solving the legal topic or things like that, and of course there's the day-to-day operation of a company that needs to be dealt with, and we're not many people. That's maybe at the more practical level. I think at the, let's call it, what keeps me occupied all the time, is that there are still places in the product that people don't enough, like, I haven't seen enough people use, and maybe that brings us back to this point about data. Like, we don't have data, okay, these are small numbers, you can't measure it, you have to look at it with your eyes, and I don't have enough, like, places where I can say with certainty, okay, this works. And that's a very, very big challenge in what we do and in the way we chose to do it.

Eyal David: [42:32] That's it, this thing of sitting, let's call it sitting on the dam all the time, right, constantly regulating basically who comes in and how much, and that's a significant challenge. I wonder whether, when the product launches, after all this dam needs to be smaller, more water passes through it, basically less control over it, right? So that's probably

First impressions and entering organizations bottom-up

Jonathan Gur-Zeev: [42:56] something you're planning, also basically planning to handle. Absolutely. The day we open it, that means we know the product is bulletproof. By the way, we can talk about how people say, if you launched and weren't embarrassed when you launched, you launched too late. Let's talk about it. I don't believe in it. First impressions are the most important thing, especially when your target audience is not people... it's not developers, it's not an API, like, with lots of patience, who love finding the problems and know, like, how to handle them. It's designers, whose cheese you're moving anyway, and you're moving their cheese very, very, very far from where they're used to it being. You can't screw up, not in the product at the technical level, bugs and so on, not in the experience, not in your onboarding and the education you do inside the product. So from my point of view, like, it'll launch when I don't feel embarrassed to launch it.

Eyal David: [44:02] Yes, sorry, there's also the matter of operations here, meaning you need to be there to hold their hand, which is also something that, with few team members, is like a challenge.

Jonathan Gur-Zeev: [44:12] Right, so again, because we know we are focused, going back to go-to-market, through individuals, through the operator, or the user lowest in our food chain. The working hands, let's call them. Exactly, the working hands are our way in, but we know we want to reach companies. Meaning, maybe I, like, maybe things will change, but today we have no plan to support endless, let's call them, Canva-like freemium users.

Eyal David: [44:46] Freemiums, in short. Exactly.

Jonathan Gur-Zeev: [44:50] And then focus the operations efforts and support and help and education really on companies and not on individuals.

Eyal David: [44:57] So one more sentence on that: how do you, how do you imagine this happening? You've brought in designers, basically, they need to basically become a mouthpiece of, hey, this tool is excellent, it saves me tons of time, like that.

Jonathan Gur-Zeev: [45:09] Yes. Let's add two more layers. One is the topic of collaboration, okay. When people work in a team, they never work alone, and like, there's no product today you'd ship, like, to a company that isn't, isn't collaboration-based. That's one thing. The second thing we discovered is that it's not only designers who have a headache when they work with AI, because there's lots, they don't know and they don't know everything we talked about earlier, but also their manager, who now needs to hold a license for Midjourney and a license for Runway and a license for this and a license for that. And you come and tell him, listen, I do everything through one license. So that's like a by-product, something that happened, and we learned it's painful for studio managers and for the managers of those designers. So it makes a lot of sense from his point of view when the working designer comes and says, listen, you can cancel all the subscriptions, get me one subscription. He'll be happy too, and yes, that's our way to spread inside an organization. I'll add to this that it connects to what I talked about at the very beginning, about the added value of AI and what it can do. So if we're talking about a million times faster or a million times more efficient, or a much more significant ROI, then for that we also developed a view for users who aren't those designers, but are the users of the machines the designer built. And then basically that's also our way to spread inside an organization. And again, you need to land on that core and have it be your advocacy, and it'll keep spreading like that. That is, the marketing people, whom we said we'd come back to, for example. Right.

Eyal David: [46:53] Gorgeous. I want to flip the record again, this time to a different place. So really, from this whole view of the career you've had, what are, in your eyes, the most important product skills of a product manager?

The product skills that matter: agency and getting it done

Jonathan Gur-Zeev: [47:09] I'll say that first and foremost, I put aside intelligence and personal talent, okay, and personal ability. I'll start for a moment with the most important thing: knowing that you don't know everything, and admitting it. It's the desire to learn and the ability to learn, obviously. Being analytical is super, super important, and on the other hand, like, a people person. A people person. You can't drive things if people don't like you, or if you don't know how to motivate them, because you're... the will to win. I don't remember who published that post about some kind of graph that says, as in, would you hire, like, someone who's super, super smart and super skilled and everything, versus the product manager with the highest agency. The kind who doesn't see anything in front of him, who'll bang his head against the wall and will manage to solve the problem.

Eyal David: [48:11] By far, I'd always choose that one. Right, it's a bit, sorry, it's a bit like the debate about whether you use AI or don't use AI. In the end, what matters, you're saying, is the end result, and in this case also that you're nice and a people person, and like, that's the truth.

Jonathan Gur-Zeev: [48:27] Exactly. I don't care if you know how to use... I'll even say that when I used to interview product managers and they'd tell me, yes, I know Jira, I know Scrum... I myself don't really know what that is, to this day I don't know what Scrum is, and I don't know what... there's another word. Anyway, it would immediately take points off the person for me. Why? Because he thinks it's important. Because the person who said that in the interview probably thinks it's important to me that he knows how to work in a certain way. And again, it comes back to methodology, it's not even methodology, it's more like processes. I don't care how you do it, I just want you to do it and bring results, I don't care how you chose to get there. So I think that's the number one trait maybe: that you get it done, like, no matter what, and you'll make up for all the other gaps. Learning professional things, you need to be able to learn, but you'll learn them.

Eyal David: [49:20] Yes, which is probably basically the parallel between an entrepreneur and a product person. I think, I'll just say about what you said, that it's what I thought: there's a lack of understanding, let's call it, among candidates, from what I hear and see. Basically there's the first hurdle, from their perspective, of, hey, I'm sending a CV and I need to even pass the first bar. At the first bar they're expected to throw out buzzwords, they throw them out, and then they don't understand that when they reach someone like you, they now need to drop that thing and really get down to substance. Absolutely. That's, that's, like, if you can take a tip for someone who's interviewing, I think Jonathan gave something very interesting here. So a last question, really around this, around the Weavy part: why Weavy, basically, why that name?

Jonathan Gur-Zeev: [50:10] A, because, okay, choosing a name for a company, there are lots and lots of parameters to take into account, right. You want, beyond it being catchy and cool, you want people to know how to spell it without you having to spell it out for them, you want a relatively short domain. But, but no, but being serious for a moment, Weavy, or weave, is the act of weaving threads together, it's what you do on a loom. I wish Loom had been available, but that's the idea: to weave together all the AI models that exist and will exist with editing tools you know and know how to work with, correspondingly adapted to the AI world, and also product-wise, meaning visually, inside Weavy's interface it's very clear why it's called Weavy.

Eyal David: [51:03] It really just reminded me of Macromedia's Dreamweaver. It's the same idea. That's for the old-timers among us.

Jonathan Gur-Zeev: [51:10] Yes, exactly. Cool, it was a really interesting conversation in my eyes. I want to ask you, basically, in five years, where do you see Weavy and Jonathan? I think we'll be a very, very, very significant part of the creation process of lots of people. Meaning it'll be Adobe, Figma, Weavy, at that seam of the layer between AI and the new world, let's call it, sorry, between the new world and the more traditional world. I think slowly, and from there, we'll expand into their worlds, but today we really... and there's some fit for that seam between, meaning we don't really compete with Adobe, but it's kind of a bit yes. So in five years we'll be a very significant player there in the creation tools of professionals in the worlds of image and video making.

Eyal David: [52:11] Maybe under the Adobe suite or something, who knows. So I'd be happy for us to come back and talk about it later, and I'd also be happy for what you described to happen, and I'd also be happy for us to talk next time about what it's like to be, what it's like to work remotely, basically, as a founder, that really intrigues me. I really enjoyed interviewing you. And how can people reach you if they want

Jonathan Gur-Zeev: [52:34] to learn more about the product or just to talk? So whoever wants to get to know the product is welcome to visit our homepage, weavy.ai, and leave details there on the waitlist. Directly to me, LinkedIn. Gorgeous, thank you very much, it was really fun. Thanks, I really enjoyed it too.

Eyal David: [52:54] Hey friends, thanks for listening. If you found this podcast valuable, you can subscribe and follow us, of course, for more episodes on Spotify, Apple Podcasts or any other app. And of course, if you didn't find us on some app, I'd be happy if you wrote to us. We'd really appreciate five stars on every platform and for you to follow us, so more listeners can discover us and find the podcast. You can also find previous episodes on any app or on the YouTube channel. We have links in the description. So until next time, come on, be effective, and bye bye.