Transcript: Integrating Gen AI into Product and Organization — with Ohad Peri (Bllink)

Eyal David: [00:00] Hey everyone, this is Eyal. You've reached Product Builder, where you can listen to the most interesting product conversations. My journey began in 2012, when I learned what product is, at my own startup. Since then I've worked in the industry, and since 2017 my company has been consulting for a range of companies. My drive is to help all of us improve as product people. And before we start, if you enjoy listening to Product Builder, please rate us five stars on Spotify. It'll give me great feedback that the podcast is giving you value, and it would make me very happy. So come on, let's begin. Ohad, how's it going? What's up? All good. Everything okay? Better than that?

Ohad Peri: [00:42] Better than that, we're thriving — I don't like saying it because it's kind of a cliché. I didn't go there. Yeah, I don't know, I said maybe you'd elaborate.

Eyal David: [00:48] But we're joined by Ohad, who comes to us as the CTO at Bllink, right? We've known each other for a while now — and how long is a while? Since the Fiverr days? Ten years, yeah. Right. We were actually just talking downstairs a bit about how this building we're recording from, used to be the Fiverr building. Really nostalgic. Yeah. And Ohad is basically here to talk about AI. This is a slightly different episode — we'll actually talk a bit about your problem space to get to know you, and then move quickly to the interesting part with the Case Studies on AI, which we managed to cover, that we discussed, and which we basically came here to talk about. And I think this is really the outcome of a few conversations we had where we said, hey, come on, this could be a great episode — and that's really the basis for this. So come on, let's run through getting to know you in a few minutes, even a minute, the way you like it. So tell us a bit about yourself — the companies you've been through, what you're doing now?

Ohad Peri: [01:44] Many years in the industry. I started in the Atuda academic-reserve program, at the Prime Minister's Office — so a bachelor's in electrical engineering, I was in the Atuda for six years, I founded a startup on my own — just me — I was the CEO and also the QA, I worked on it for eight months, all full time, eventually, my bank account signaled to me that it was time to move on, and that's it — that's how I got to Fiverr.

Eyal David: [02:12] A great school for startups, an excellent company, big and growing.

Ohad Peri: [02:20] So, from there I moved on to HealthyIO, where I managed the entire back-end, all of their infrastructure, also an excellent company that developed a medical-testing product that passed FDA approval. From there I moved to a very successful, global company that's publicly traded here, which basically rates Bitcoin, and from there I became part of the founding team at my current company, which is Bllink, a startup dealing with building-committee (va'ad bayit) payments. Okay, so tell us about building-committee payments,

Eyal David: [02:58] give us the short version, and then we'll maybe dive into the AI aspect of things. So, building-committee payments in Israel — many people don't know this, but it's a huge market,

Ohad Peri: [03:07] 15 million shekels a year. Wow. Really, huh? Think about how many buildings there are in Israel.

Eyal David: [03:15] I wonder how many of them pay a building committee, but yeah, okay. Everyone pays, it's mandatory, it's mandatory in Israel.

Ohad Peri: [03:22] And if you go back five years, 90 percent of them paid by check, which is a very big problem — a logistical, operational problem for the management companies, and we save them — we basically solve the problem with a web interface, where the tenants can pay easily, sign in once and forget about the building committee for life, or until they move apartments. Yeah, you can set up a recurring payment for the building committee, and the management company doesn't have to deal with payments.

Eyal David: [03:54] By the way, in my building there's a different company — it's not you, funnily enough. Yeah, we have competitors, yeah. It's not good. It's not good. So where's the connection to the problem here — where did you identify it, as someone from the founding team, like, how did it happen?

Ohad Peri: [04:07] So, a little under a year ago I also took on responsibility for the product, and I came to the conclusion that we needed to bring AI into the system already.

Eyal David: [04:17] From which angles? Because it was a tough year, right? Like, efficiency — it came from there. Exactly from there — exactly from the efficiency drive; we also want to become more efficient in R&D,

Ohad Peri: [04:27] and also to introduce features that would come up, that would automate things. So we started with a first feature that ultimately — a little curiosity — didn't succeed, but building it brought the AI infrastructure into Bllink. So wait, we need to explain what it did. Yeah, the first feature — the thinking behind it was that a building committee wants to onboard a new building. They'd be able to do it through AI, in a chat, without needing a human rep. So we built it. It took us a month — a very long time for a startup — and in the end, that assumption turned out to be wrong, but we gained an infrastructure out of it — an AI infrastructure that we use today, and we use it mainly to automate customer support. Many of our managers like to ask our chat questions, and that's how we gained — we gained efficiency from it.

Eyal David: [05:26] And you also learned a lot, so I have to dwell on this. Yeah. So it's really intriguing, because it was actually the psychological aspect that failed, because in the end people — building-committee managers — ultimately expect to meet, like, a person who picks up the phone and sets it up for them — how does it work?

Ohad Peri: [05:40] Yeah, the moment you do an onboarding that's a bit complex, you need some person to reassure you. That it's fine — in a product that deals with money, that there won't be some situation where you don't get paid or something like that,

Eyal David: [05:57] so you want to see there's a person behind this thing. So that was the wrong assumption that... anyway, I actually know another company, we were just talking about Lemonade a second ago, right? They do exactly this — an insurance company — and they basically created a super-easy onboarding, and Lemonade's advantage is really that they have the budget — the budgets to really brand themselves, as a player with very high trust. Exactly — to build up the trust and a very relaxed experience. Okay, so that's apparently the other side of the coin, which you need to consider when designing something like this. Okay, so you learned, and with you... want to say a couple of words about the infrastructure you built? Is it a screen? Yeah. So, when you built the infrastructure,

Ohad Peri: [06:39] all kinds of solutions presented themselves as options, and in the end we chose a modular infrastructure. We split it into three. One infrastructure is the chat layer, which isn't actually related to AI — just chat... Conversation. Yeah, a third-party we used. The second infrastructure is a set of actions we can run in the system. And the third infrastructure is simply the AI. That is, given a certain sentence and a certain context, the AI decodes the user's intent, and it knows to run the appropriate function with that user's permissions. Wonderful. Okay, so that's clear — that's your layer, essentially.

Eyal David: [07:26] Now you're basically saying, I can detach this part, this puzzle piece of the GenAI — which right now is GPT, I don't know what you use — and you can swap in something else, while all the data you collected is still kept in your layer, the logic is still kept at your level, and it makes the decisions. Right, exactly. So basically, in the end,

Ohad Peri: [07:46] we discovered that this infrastructure is actually really simple. It genuinely gave us scalability, not just in the sense of speed, of a very good product that the managers love to play with, but also very high speed — think of it this way.

Eyal David: [08:05] You don't need the front-end, right? Ultimately, you simplified it completely.

Ohad Peri: [08:12] You simplified it completely — it's a chat, right? The back-end logic too. Once you've implemented the function you already have, probably, in full, unless you need to build it. You only need to write some specific text: given that the user asks you about X, do Y, and the AI is really good at decoding that X. That is, if he asks it from this angle, from that angle, it'll know he means X. You know what else is lovely about this whole story?

Eyal David: [08:43] Presumably, even your own layer you can build with the help of AI — both the layer and the functionality, built that way,

Ohad Peri: [08:51] or maintained that way. We definitely use GPT to code, Copilot, and recently there's also a new tool, which we'll also use, called Cursor, which is really an IDE. An IDE, really from scratch, that uses whichever AI model you choose, to write code. There's hype around it and we'll pick it too, so we use AI in all the layers of the product, so to speak.

Eyal David: [09:19] Wonderful. So this whole intro brought us to a place where we understand why you're here, why you've gotten so deeply into the field, why you publish a lot of content, really good stuff on LinkedIn — which I'm exposed to too — around the field. So that's why you're here. And so, let's dive in, let's talk about all kinds of examples of AI uses. You gave an example, for instance, about the building-management segment, there may also be uses for coding, right? To boost efficiency? Which other segments do you want to address,

Ohad Peri: [09:47] across industries? This segment of coding with AI is actually the most interesting to me, both because it's my profession, and because developers are really the driver of productivity in all the other areas of the market. So the moment you raised developers' productivity, the increase directly translated into market value too, because the features ship at a faster pace, and one of the metrics is the 50% increase in Time to Merge. Time to Merge — meaning the code goes to production, or it's merged into the code of... 50% faster, amazing numbers. Use Case 1, is a company called Vercel — they developed V0, I use it quite a bit, and basically, in a few minutes, you can build yourself a front-end, one that even works, even with JavaScript logic, in English. So like, I... What do I actually give it? I give it a prompt, or I give it an image of the interface — see? Let me give you a really real example of our own, I wanted a mobile-optimized menu for bank transfers, that our managers want to see — a report, and I wrote it out to it in English. It creates it for me. Of course there was a V0, V1, V2, and eventually we settled on V8, but, you know... so, like, just to tighten it up,

Eyal David: [11:09] so you get something that isn't designed — well, it's designed up to a point, right? Because, like, I don't give it any design,

Ohad Peri: [11:15] and based on that it generates your convention. Yeah, but it has best practices. Yeah, exactly. Yeah, it has best practices, how to build something mobile-optimized — it knows, it's familiar with all the terminology, the code, all the design lingo, of design and code. Exactly — it takes all the heavy lifting, and then it basically leaves you with just the aspect of 'come on, dress me up however you want with design and fonts.' Totally, and in the end, I gave it our glossary — Bllink's — our theme,

Eyal David: [11:42] and it applied it onto the product. Wonderful — how long have you been working with this? With V0?

Ohad Peri: [11:48] We've been working with all kinds of other tools for a few months now, with Copilot for a year already. So you brought this in a while ago,

Eyal David: [11:55] all this goodness, all these tools — how long ago? Something like a year, yeah. And do you know, as the CTO, how much it saved you?

Ohad Peri: [12:03] In time to merge, or — I don't know if you measured it, it's really intriguing. We didn't measure it, because in a small startup there's a saying I like to use, measuring efficiency is terribly inefficient. Yeah, okay, so you just run with it. So yeah, we run with it, but we see the efficiency of it.

Eyal David: [12:18] How nice. Cool. Something you could maybe pass on to companies too — slightly larger companies — about how to adopt AI — it sounds really interesting — in the development process. Where else did you come across things, like, dev tools, besides this tool?

Ohad Peri: [12:34] Another product we haven't used, but we see it across the industry, is a product that basically replaces your DevOps work — so, companies that are hosted on AWS, you can use a chat that knows the context of your environment — 'environment' meaning in the cloud, I should say... Yeah, it knows all the instances you have, it knows all your network connections, and you can literally present it a bug, and it'll present you the solution, given your actual environment. Or it can also recommend upgrading, within AWS, right?

Eyal David: [13:10] like taking the action, or just advising on the whole process, right? To upgrade the... yeah, I don't know, how to move another parameter in your environment, like fixing it myself, I don't know. Yeah, there are other tools that are basically FinOps AI too,

Ohad Peri: [13:24] for saving costs — so they can recommend swapping an instance, upgrading, downsizing,

Eyal David: [13:30] yeah, honestly that's not really in AWS's interest, and I think about that. Interesting — do you want to say a bit about the Sequoia research? We'll of course put links to all these things, neatly organized with the episode, but do you want to talk about the Sequoia research,

Ohad Peri: [13:44] some info that likes the environment? Yeah, so at the macro level of this trend, if we take a step back, we see this is a very new technology, that costs a ton, a ton of money to develop — to train the models, especially the Foundation Models, and Sequoia published an article — Sequoia the VC, one of the biggest in the world — they published an article, 'The $600 Billion Question,' which is basically the amount that had been invested by then, a few months ago — so by now it's surely the trillion-dollar question, invested in AI startups. And I can't imagine what the criterion was there, right, like whether it was ever a company tagged as AI, or from a certain point, and what's the measure for saying a company is AI-focused? I don't know, I think the criterion is really companies that develop AI themselves, and don't just use it — though I'm not sure — but in the end, it's a very, very large investment, in developing this technology, and the question they're asking there is basically, is it worth it? That is, there's currently no justification for this investment. Anyway, I've been asked this question in all sorts of places, my answer, the way I responded to it, was that I looked at a groundbreaking technology that happened 60 years ago, in the build-out of the electrical grid infrastructure in the United States — I did some research, and the ratio of GDP, in the costs there, was 10% of U.S. GDP. Meanwhile, 600 billion is less than 10% of U.S. GDP. So there's room to scale the story, this expenditure? Yeah, and another study, at the macro level, from McKinsey, forecasts an increase in global GDP, at an annual amount of 2.6 to 4.4 trillion dollars, an addition to global GDP, which currently stands at around 100 trillion — so it's basically an addition of 2% to 4%, just for comparison — yeah, London's GDP is less than that, England's. So let's really dig in for a second,

Eyal David: [15:58] before we move to the next case study — what does it actually mean? We took the segment of doing development, okay? I'm now making things more efficient, which means I'm shifting people, basically away from what they do today — a computer will do it, AI will do it, efficiency will grow, because now, for that matter, as a solo entrepreneur, I can work with just AI — with AI tools — and give up the team I once had there in the warehouse, and run a company — there are tons of stories like that on LinkedIn. Right. It's really interesting, because suddenly you shift the entire middle, you and I talked about this before too — the whole middle, you developers, who on one hand will suddenly be weaker, development, developers, QA — everyone in the company who was in the middle basically shifts, AI replaces them, efficiency rises, GDP rises, so interesting processes are happening here, and you and I talked about this too, that maybe one of the signs of this is that even the time it takes to improve as a developer — from mediocre to good — probably shortens too, because the work tools and the learning tools are faster as well — it's interesting how else this will impact, from this efficiency, the job market — what people will actually do with their time. Maybe, Ohad, you're a good example of this — when we talked in the pre-interview,

Ohad Peri: [17:09] about more products, more problems — so first of all, honestly, I don't know what this technology will bring to the world, some say dystopian — meaning very massive unemployment, it'll replace all of us, and it won't even be worthwhile for companies to employ humans — but a more optimistic vision, that is, increasing the number of problems the world solves — so if right now we're not focusing on space exploration, because it's too expensive, or we're not focusing on studying our own bodies, because it's too expensive, then maybe solo individuals will solve the problems that huge, giant companies solve today, those problems, and giant companies will solve the problems that no one solves right now, so maybe that will be the — that will be the solution to unemployment in the future.

Eyal David: [18:01] That was a very interesting take on this. Okay, let's move to the next case study. So we actually wanted to talk about the automotive industry. Really interesting, right? These are exactly the industries where every such technological innovation impacts their efficiency. I don't think there's been such a technological innovation, since the invention of the computer, that so impacted, perhaps, the entire heavy industry, right, the revolution that happened before our era. So come on, tell us what happened at Toyota, it's really interesting.

Ohad Peri: [18:33] So if we really drill down to see another case, we see that this technology is adopted in companies where the regulation is lean, yeah, there isn't much regulation there — design, development, etc., etc., etc. And in industries with very heavy regulation, like healthcare, for example, and like government organizations, it's very, very hard to introduce the new technology, which is very, very unpredictable. One of the industries that did bring it in, despite there being a safety element that has to be overcome, is Toyota — a company very well known for its efficiency, they introduced this tool they developed to generate futuristic designs for the new cars — basically the tool, given a certain text, can generate design variations, so the designer can basically tell this tool, generate me some futuristic variations for this-and-that car, and produce minimal drag — a car's drag. Minimal car drag — what do you mean? basically that the aerodynamics are high — that the car, as much as possible, consumes as little fuel as possible to move forward. And on the face of it, I could presumably also create a model that checks the car's safety, and other models that would probably render redundant the processes that... okay, they need to be done, sure, but... making the cut, and maybe tossing out the designs that won't make it — so I can do it much faster, and then test much better variations. So that's exactly what they did, the optimization for X, where in this case X is the car's aerodynamics — so the same thing can be done with Y, the car's safety, to maybe skip over the simulators they use. There they probably won't do it, because safety is something that...

Eyal David: [20:32] Yeah, but you can probably filter — you can filter out the ones that are probably... like, out of the mass of designs that suddenly get created, because far more designs are created, you can filter from there and end up with a mass that's more, you know, optimized for safety. It's really surprising that the Japanese do this, right? It's a very Japanese thing, to advance and do things like robots — it's interesting that...

Ohad Peri: [20:55] the designers sit in California, so apparently...

Eyal David: [21:00] Cool — I can think of many more industries; say the construction industry could probably make a lot of use of AI, or not, as a case study, but many such heavy industries could really make use of AI — for planning, for analysis, right? These were basically the main cases. I'm sure there are more — yeah, we just didn't find them, yeah. We didn't search enough, right? It didn't interest us enough. Let's give examples, most of them from our own world — so we talked about software development; a moment before that there's also design, photography,

Ohad Peri: [21:29] Adobe's AI tools — do you want to say a bit about that? So yeah, Adobe is a great company that took a bit of a hit from Figma in recent years, and to compete with that, it too introduced a Gen AI tool of its own — Figma introduced one too — which, given a text prompt, like we see in V0, lets you generate the design — in my opinion, it's actually a waste of time, the designer — if he can also generate the component together with the code,

Eyal David: [22:03] then it's not worth it for him to use Adobe or Figma. Really — I think it's interesting how far they'll take this, at least at first, those designers, it's really easy to take it — okay, come on, I'll do the basic design and then you stitch me all the edge cases, all the variations, and that's classic. Yeah. Very quickly — I think I used it early on, you know, to generate design concepts, with — early on — Gen AI, I have images, interesting, but where will this take us? Because very quickly people are already talking about many such industries going extinct, or, as you say, shrinking in terms of headcount — it's really interesting what will happen to the UX market and the design market, right?

Ohad Peri: [22:39] Yeah, I can see companies getting punctured by this — for example, the company where we worked, Fiverr. So Fiverr — I don't know, from the inside — but Fiverr is in trouble right now, because its whole basis relies on freelancers who do design. You know, at our company too — a year ago we had a freelancer on the team that we used to use.

Eyal David: [23:03] And now we use her less, for internal projects. This aspect is actually really interesting, because you have to think about it. If I now have a company with a certain design language, I don't know — for me at least, using all these tools, it's hard for me to keep design consistency between the things I generate,

Ohad Peri: [23:21] everything always comes out different from one another. So for example, in this tool — when we talked about V0 — you can load your theme, basically your design language.

Eyal David: [23:29] Exactly. So I need to come with that thing — it's interesting, it's interesting how long it'll actually take for this whole long chain to happen, all the way from, like, generating a unique design language to actually working code — it'll probably take a few years, like, until it's...

Ohad Peri: [23:45] And not necessarily — I think it won't make designers redundant, at least not in the first stage, that is, ultimately, it makes redundant — the famous saying — it doesn't make people redundant, it makes redundant the people who don't use this technology. So a designer who uses this technology can simply do, instead of one job, three jobs in the same amount of time.

Eyal David: [24:06] Exactly — I think what you said about solving far more problems is really interesting. That's probably where this is heading,

Ohad Peri: [24:12] it's probably good news for humanity. I can actually expand on Adobe — they also have the option, again, in the photography world; my wife is a photographer, so I know that completing images, swapping backgrounds, with all the good things happening here too, and even in simpler tools than Adobe's, which are for more professional photographers — it's pretty amazing. And I think in video too — all the videos created by AI recently, you see it everywhere, it's also very amazing, and if you haven't seen it, it's really worth going in and checking it out. Right — we didn't give use cases, but the creative industry,

Eyal David: [24:44] and film creation, marketing, sales, all the asset creation for these things — it's completely undergoing disruption. Again, it's not there yet — because in our own experience we tried to generate, our salesperson tried to generate, assets of himself to save on that, to do it overnight — it's not there yet, but it's getting close, it's totally an industry undergoing disruption.

Ohad Peri: [25:09] Maybe — I think if you're an enterprise, and you have access to enterprise-grade tools, then maybe things look a bit different too. I just read about this recently, that people are afraid politicians will be replaced by it, because you can suddenly generate a video, you can generate a narrative — a narrative that responds to and fires up the audience in a certain way,

Eyal David: [25:26] and, like, generates power, generates... well, I have news for you. Tell me, tell me. It already happened in London, in the last elections, they created an AI candidate — based on a real person — and ultimately he created a sort of clone of himself, which he loaded with all the values he believes in, his platform — basically an AI that talks with real voters, and can handle, instead of that person — that candidate — the conversation; that person is limited in his time, AI isn't limited in its time. Wow, that's just amazing. I'm glad I brought it up. It's also really interesting, again, about the animation world, many worlds are really going to undergo serious disruption, and it's right to focus on the development world — it's really interesting to know where this will go. Can you tell us in a few more words just about this IDE? Because it really... I hadn't heard of it before and it's really interesting. Ultimately I assume most viewers are familiar with the productivity tool

Ohad Peri: [26:28] that broke into the market, called GitHub Copilot, which basically gives you — I use it — gives you a really, really, really, really good autocomplete — basically, given your context, it knows the writing style in your codebase, and it gives you really, really good completions, so this takes it one step further. Basically, if you want to combine this tool called VFS, where you basically give it a very, very high-level prompt of what you want it to build, and it gives you fairly complete, ready-made components,

Eyal David: [27:07] and even does bug fixes and code changes. So it's simply built into the IDE, it's basically...

Ohad Peri: [27:13] Yeah, you open a prompt window and enter it, and you can also choose which files go into the context, because you can say, given this file I have, and this file I have, go build some third thing that's based on these two things, and all in free language. And it should be said that now you can also write, presumably, all the Stories and all the Epics, with the help of AI too,

Eyal David: [27:37] so here and there we'll see where we're left in this respect. Do you want to mention, a second before we finish — because we ran a bit fast — which other tools you use at your company that might be worth adopting, that you can expand on? We use DALL·E,

Ohad Peri: [27:53] to generate creatives, basically. Yeah, does it work well? Yeah, works great — we also run all kinds of ads at the company for the things themselves, so yeah, we use it to generate assets for ads. Okay, for optimizing your campaigns,

Eyal David: [28:13] do you also do... we didn't get to talk about Shopify Magic, so maybe let's touch on it from this angle — do you use some one-click thing for marketing, for performance? No. An agency, like in the good old days? Yeah, yeah, yeah. Cool. For now. Exactly — whoever's listening to us. So also, your tasks — where do you manage them? The tasks of? Your tasks, generally. For instance, I work in Notion, so it's Notion — the AI works really well. We're on Jira. And in Jira too — which one? The AI is strong, right? Do you use it? No. At all? No. Well, so there's an option there too? No — there too, it lives, you know, it lives inside the software.

Ohad Peri: [28:50] What, for the description of the ticket itself? Yeah — what's interesting, what I'd expect to reach the next level, is that... okay, here are the features you're suggesting to me. Come on, give me their benefits already, give me the impact you think they'll deliver, arrange it for me, look, soon. And also decide for me, basically, how it fits into the scoping, how I prioritize things — maybe I'm not even on the right track, and there are questions I need to ask,

Eyal David: [29:12] like, that too. Really, I hadn't thought about it, but yeah, one of the... maybe it's an interesting topic to talk about, not in the context of AI, in the context of product — I really... I came to the right place. I took on this role not long ago, and I listened to another episode, with Yael Shamir, whom we know, and she said a line I really agree with — that it's really like, you feel different muscles, having moved, you know, to managing development, and managing the product too, you really feel different muscles in your brain working, and you ask questions you didn't ask before, about why to do this, how much impact it'll have on the business — and I think, absolutely, I can help with these things too. Right — again, models, and it should also know the company and everything. So this seems like something worth taking on — so maybe take it on and come back to tell us how it went. I publish regularly, so you just need to follow. Wonderful — so where can people actually find you, if they want? I'm on LinkedIn, on Twitter. Also, it's important to say that Ohad would be happy for people to reach out to him with questions around this topic, around decisions on the processes, how to use the tools. Gladly. Wonderful. And we have one more question I want to ask you. Uh... you actually mentioned — usually I ask this early in the interview — you've worked with a ton of products at many companies, you listed them all, what, in your opinion, are the most important PM skills a product manager needs?

Ohad Peri: [30:48] I'd say that in first place it's transparency. I've often run into some communication gap between product and developers, where the developer didn't actually know why he was doing what he was doing, or the logic that led to the decision on this feature. So I'd say, good communication — breaking down the logic, high-level: what happened that led us to work on this feature? First place. Second place — PM skills, for which you need — this I also discovered when I came in a year ago — you need to say no a lot of times, because there are many requests — more than any company's capacity — because ultimately resources are always limited. You have to say no, and when you say no, you need to know why you're saying no, you need to explain it well, and you need to explain it nicely too.

Eyal David: [31:41] If we think now, in a time machine, ten years from now — where do you see yourself, what do you wish for him? So ten years from now I'd be happy to be doing what I'm doing now,

Ohad Peri: [31:51] basically a technology leader — this position has been very critical, between business, product, and technology, and I think it's important that every company has this X — a person who knows both the team's constraints, what they're good at, and also knows what needs to go to market, and does this X, and knows how to make the best decisions, and I'd be happy to keep doing this ten years from now too. Wonderful — it's purposeful, it's really interesting to do this, so lots of luck with it, Ohad, thank you so much for coming,

Eyal David: [32:23] and follow Ohad on LinkedIn, it'll be very interesting, and until next time, thank you very much. Thank you very much, my pleasure. Alright, bye for now. Hey everyone, thanks for listening. If you found this podcast valuable, you can subscribe, follow us, for more episodes of course, on Spotify, Apple Podcasts, or any other app — and of course, if you didn't find us on some app, I'd be glad if you'd write to us. We'd love 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 the previous episodes on any app or on the YouTube channel, we have links in the description. Until next time, come on, be efficient, and bye-bye.