Bananaz and Jiga - Design Faster. Source Smarter. Ship Sooner!

Michael Finocchiaro: And we're live. Thank you for joining us for AI across the product life cycle. I'm talking to two interesting companies today, both in Israel. I have Adar Hey of ⁓ Jiga and Or Israel of Bananaz Welcome guys.

Adar: Thank you. Or Israel: Hey, thank you very much. Happy to be here. Michael Finocchiaro: Why don't you guys introduce yourselves and your companies real quick so that people know where you're coming from.

Or Israel: So I'm Or Israel the CEO and co-founder of Bananaz ⁓ My background is in mechanical engineering. was head of engineering at DPFR and for a medical devices company for about six years before founding the company. And I'm actually dealing with CAD world since like in the development side of things. ⁓ In we're building a gigantic layer that sits on top of CAD software, such as SolidWorks 3 and so on.

And we help to save time and money by creating agents that automate manual tasks and helping them to be way more productive. Michael Finocchiaro: Awesome. So you're more on the design side because I think Adaris is more on the supply side. So go ahead.

Or Israel: All right. Adar: My, that's right. So I'm Adar CEO and co-founder of Jiga. ⁓ I coming from not an background, not a manufacturing background, actually a marketing and sales background.

⁓ And we just set out to solve this problem after we seen how hard it is source a custom parts and how inefficient it is. ⁓ And we are building a... ⁓ system that helps companies source parts more efficiently, more reliably. We basically plan within the system relationships with custom parts manufacturers and software in our own operation to help them get parts reliably on time, et cetera.

And we use a lot of technology on top to help that happen. Yeah, that's it. Michael Finocchiaro: Okay, it's good we're talking about agentic stuff. I've got a new sponsor AWS and there'll be a URL at the end where you can download their ebook about agentic architectures and best practice.

think there's 15 or 20 experts in the book. I'll put the link in later. So if we look back to 2022 when this whole crazy AI journey started for all of us, What ⁓ were you guys like skeptical about the possibilities of AI or you guys super bullish from day one about AI? Adar: I think when I saw it, it was an incredibly impressive moment, I guess, for most people, because we haven't seen anything like that.

⁓ So definitely got me curious. think that as time passed, I started seeing people getting, you know, I have been skeptical, especially because the technology wasn't anywhere where it is today. but I have seen the potential where it can go. So I would say a mix.

I would say very bullish, but just skeptical about certain hype factors that were there at the time. Michael Finocchiaro: Certainly the hallucinations too, And in order, you're an engineer like me, so you like determinism. So I suppose it was a bit disquieting because of all the probabilistic aspects of it. Adar: Yes.

Or Israel: Definitely. think that, ⁓ you know, this is a question that me and my brother Ophir, who is my co-founder and our CDO, this is a question that we are talking about like ⁓ quite a lot. Back in the time, this was like actually ⁓ like not as new as for the majority of people because Ophir was like working in Amazon and back in the time they were dealing with LLMs and with AI for years. It's just like wasn't like in the in the same kind of LLM models that we have in Chagy Pt and so on and if you ask over here about that AI is one solution out of a portfolio of solutions when you're dealing with technology and with like trying solving problems and where I came from I tried to solve my own problems and my problem was that we had inconsistency in the reviewing drawings and reviewing models and In that case, AI is not the best technology to solve this kind of problem.

The problem needs 100 % of accuracy. You need to be able to go and track changes and make sure that nothing is being overlooked, especially in medical devices. AI gives us lots of capabilities in various areas. And we have tons of features that are based on AI.

But the core thing that we did three years ago was extracting data from drones and models. And in that case, ⁓ that's not the right technology to use. On the other hand, AI could give you so many like value in some other areas in which you could get either insights, suggestions, things that are not deterministic. And therefore it's like an incredible technology that both like getting so fast, so good.

like, it's kind of crazy. And I think that the capabilities are in the pace of development becomes like significant. I think that's something that we feel that is super helpful. Michael Finocchiaro: Yeah, in fact, that segues into our next question, is ⁓ how has ⁓ AI changed the way you guys develop software ⁓ at R?

I suppose like everybody else, you've been using a cursor or cloud code or whatever, ⁓ how ⁓ is AI influenced the way you guys developing software? Adar: think that AI is a fundamental change in ⁓ just how products are being used. Because suddenly, software is not necessarily only a tool for humans to use. It's actually a means to get work done.

So that changes your whole perspective on how you build products. And whether even you need all your product to be in the same Or Israel: Thanks for watching. Adar: I think like behind the same logging screen, right? ⁓ So who said that the product should be somewhere?

It could be served to you wherever you want it to be served, right? So suddenly, like you start speaking in outcomes and less about features and less about seeds and more about consumption. ⁓ So, we're thinking about the product and how we build the product. That becomes a core thing we're thinking of is like, how can we deliver outcomes?

Not necessarily even have the users use the product, right? If they, like, eventually they need to have a certain goal in their mind. If we can help achieve their goals with even less clicks or time or logins or whatever. Or Israel: Thanks for watching.

⁓ Adar: That's a good thing, right? If they can stay with their email or whatever they're using and just have that work get done in the background, that's a good thing. So suddenly you change your whole perspective on how products are getting built and that influences our product decisions for sure. Michael Finocchiaro: And or I suppose everybody in Bananaz must be on cloud code by now.

Or Israel: ⁓ yeah, for sure. Like, ⁓ it's, it's up until like the, security here, secretary here. It's like everybody's like in, in actually in cloud card. And I think that it affects not only the product and the R &D, which for us, it's like kind of the obvious, like it's something that, ⁓ that is getting better and better like every day.

And it's like across the whole platform and whole development chain. But I think that you could get. Lots of value also in the interoperability between the different teams, which is kind of amazing. how fast you can get feedback from customers and create features out of that and let them try that and understand what's interesting for them and what's not.

And now all of the cycle of development becomes both like sharper, but way faster. We can see that like in the last two and a half years. things were like, it's escalated like, I think at least five to 10 times. And you could see that when you create the software and get like feedbacks, like, know, in startup world, you want to go and try stuff and then like get the feedback and iterate fast.

Now you could see that there are like deployments two or three times a week, which is, it's like, it's phenomenal. So yeah, I think that that Cloud and you know, before Cloud there was like cursor and you know, like shifts all the time. This is like one of the things that we see and understand that, you know, it is not only like the type of software every once in a while, there are going to be like a newer model that's going to give you better results. We do it like these two layers.

One of them is to go and take this like kind of mentality and attitude and make it like the company's DNA that you need to go and find the best tools. and implement it in every process in the organization. ⁓ But on the other hand, it's also like FXSZR said on the customer experience and the experience that you want to bring to the customers that they would feel the same kind of experience that programmers has, but in their ecosystem and space. So we want them to be able to choose the latest models and to be able to Michael Finocchiaro: Thank Or Israel: go and improve together with the big ⁓ models like ⁓ providers, open AI and so on.

Michael Finocchiaro: So. So like, it makes me want to ask like, how are you managing it for a cost point of view? Because, ⁓ you know, I've built stuff with cloud code and I don't know how much money I've spent, maybe four or 500 euros this month already. I'm just trying to update my website and stuff.

How do you guys manage that? Because if you've got a whole team of developers that are spending 1500 bucks a month, your burn rate is going to be rather high. Isn't it as a startup? I mean, your cost between the cloud costs to AWS and Azure where you're hosting or Wherever plus the anthropic cost.

How do you keep that down? Adar: I mean, the savings are just like when I'm looking at costs, I'm usually looking at what's the alternative and what's the opportunity cost. I'm honestly just looking at it from a cost perspective. And I think that ⁓ what you can achieve with the same amount of humans, right, is so phenomenal just by ⁓ paying those credits.

So every... person on our team, not necessarily the developer, but everyone has ⁓ their cloud account and they have their ability to spend. If they're asking for more credits, they get it instantly. So the way I'm thinking about it is like, they're using it, they're just deploying work at their fingertips.

is hopefully like high quality work and it is much much cheaper than hiring additional humans for example so yeah my perspective here is that the ROI is phenomenal and not necessarily as a you know only the cost side of it. Michael Finocchiaro: Yeah. And so do you have agents showing up for your Scrum meeting said Bananaz? Or Israel: Definitely.

Now I completely agree with that. It's like the ROI is like no brainer, like for every position, not only like the developers, every position that you have in the company, you need to think ⁓ heavily if you want it to be like a new hire or you want it to be a combination of agents and team members to be more effective. You can see that with way smaller team, you could achieve way more because the communication between people becomes like the bottleneck and not the, the, ⁓ their productivity in, creating code or creating their like, ⁓ kind of work.

want to have more experienced people that could deliver more in, in like using agents and, and, and using the latest tools. I think this is also something that we are trying to like sell to our customers as well. Right. Like in our case.

at the beginning of the road is selling productivity and selling them that they could do more and they could eliminate manual tasks. This was a challenge. Like when I went to my first fundraise, like cloud code was only like, you know, maybe like in the lab, but we wanted to do that. We wanted to make the engineers more productive.

And it is. changing the way that people consume their products, right? It's not only like SaaS, but it's also based on tokens and consumption. And this kind of shift of a mindset for us at least, it's super helpful.

It helps to educate the market towards like how things are being done in like 26. Thank Michael Finocchiaro: It seems it's just changing so quickly. I wonder too, in terms of the actual product, so Jiga and Bananaz, ⁓ where is AI sitting? suppose particularly with Bananaz, you have a co-pilot.

So AI is right there in the user interface right away. ⁓ But I suppose it must be permeating through other parts of the stack. ⁓ Where is AI sit? Is it part of the DNA of the product?

Is it? The maybe the thing you're having a foundational model or is it just on the user interface? So I wanted to ask both of you guys where I sits in your stack, basically. Or Israel: So in our software, it's like it's basically embedded in many areas.

It starts from the user experience using the chatbot because eventually you want to talk with your design and you want to be able to go and use your queries. But we also use lots of agents that goes and perform tasks for the engineers that run and uses different algorithms and tools to solve unique problems. that AI is not the best solution. Like in some cases, AI is phenomenal.

Finding information, getting ⁓ lots of data and give you precise answers. Go search over the internet and find ⁓ like standards or ports or anything like that. This could be ⁓ the tech that will do. Everything that based on text, obviously is something that AI is phenomenal at.

But sometimes as we said earlier, you want to be precise and then you need to go and run agents to perform tasks either in other systems, other algorithms. And it's basically a combination of both. Michael Finocchiaro: At ours is same same with Giga or maybe it's slightly different Adar: So, GIGAB, we are delivering the work, we're delivering the parts, it's fundamentally like a service that is enabled by AI in many ways. So, we're thinking about it from the perspective of like...

Our customers want the business outcome. want the parts. They want the high quality, good lead times. Cost can be a major consideration.

And of course, wasting as little time and attention on doing so. so we're thinking about AI as means to get to these results. So there are many. places in the sourcing of custom parts process that is high friction, that is highly inefficient, like things can go wrong.

So we're trying to put AI where it matters to these outcomes. So for example, identifying a potential delay or identifying a mis- documentation miss and fixing it. ⁓ on the spot without even the customer knowing, right? It delivers better outcomes to the customer or matching a need to a supplier in a better way can reduce costs and can get better lead times and capacity, right?

So these types of specific use cases where they directly tie to business value. ⁓ is where we want to plug AI and sometimes it's customer facing, sometimes it's supplier facing, sometimes it's team facing, right? But the overall arch there is how do we become better partners for our customers. Michael Finocchiaro: And in both cases, are you guys leveraging one of the off the shelf, the OpenAI, Anthropic, whatever, or are you guys using some of the open source models like Quinn or Nematron or some mix of the two?

obviously you need to... Well, I think the question is going to come up more more. How do you deal with the ongoing costs to the customer of the tokens they're using, right? If you give them unlimited, then you're going to go broke because your bill to Anthropic will be...

Massive. So how are you dealing with that to make it useful and yet not breaking the bank and having to ask for more funding every three minutes because you just ⁓ you can't even pay your anthropic bill. Adar: Yes, so our kind of calculation is whether our customers are, ⁓ you know, we're giving them value and we look at how much value we get from them, like how much valuable are they for you guys' business. And I think the same answer on the team applies here in terms of cost, because if we eventually see the retention that comes from customer expansion that comes from customers, and the value we give them or even give to ourselves by not having to put humans to validate things or to have less errors or have worse customer experience, So again, the cost is negligible compared to what we used to do.

And we have a good benchmark. We used to do things a certain way, and now we suddenly have a new way of things to do. yes, credits are the cost. But compared to what they used to cost to us, it's negligible, at least for us.

Michael Finocchiaro: can a user bring their own OpenAI or Anthropic Key or you're managing all of the ⁓ AI on your end? Adar: We manage our own AI and we connect to the known LLMs. Michael Finocchiaro: Okay. Is that the same for Bananaz?

Or Israel: Actually, I'll split the answer into two. So for the one that you just said, we allow people to bring their own models and to connect with their own keys. We have like two different motions of companies that we're working with. We're working with Fortune 500 companies in which this kind of bringing your own model or your own like contract with the big vendors is something that is pretty common.

And there are like the motion of midsize enterprises in which they are asking to get everything from you. So we provide like both options and in the matter of the usage base, think that's an important one. especially as of today's price of tokens and how fast tokens could disappear. ⁓ I would go and say like, ⁓ repeat on what Adar said.

Eventually, we want to be able to go and save money for the company so it will be a no-brainer. So for us, for example, if to create a return report costs, I don't know, an hour of an engineer, which costs about, I don't know, $60, $70, and we could provide the same kind of value in 50 cents, it becomes a no-brainer. And then, like people, like... they want to pay the money because eventually it saves them time.

And this is the kind of impact that we have in such cases. Together with that, we created our own unreadable formats that makes the process of using the tokens way more effective. it's analyzing an image. It's costly, right?

Like the formats are costly and it's very expensive to go and use that. But when... we create our own formats, it's become relatively cheap because we know how to index that in a way that it will be effective with the LLMs. ⁓ Michael Finocchiaro: Maybe another way of asking the question is what about security, right?

Because the issue, another issue with the open LLMs is of course, they're training on everything you give them, whether they tell you or not. Remember Anthropic had lied to us in the beginning saying they were never going to do it and actually they were doing it and they had to admit it. So how, if you're using the commercial ones, how are you guaranteeing that the customer proprietary, the IP is not leaking to to enter the open source stuff. Or Israel: So first of all, ⁓ we are not moving all the information to the LLMs as is.

I think that making encryptions and providing all the relevant data, this is the key to make sure that you're standing with the security of your customer. I think that's our number one priority. So when you provide text without the context, without the image, then basically it's worthless for the for the LLM without the context of the image. And when you split it into two, you get like way higher security in that aspect.

In our case, we are SOC 2 certified ⁓ since day one. We like put lots of efforts on the processes and making sure that everything is highly secured just because our initial customers were very big enterprises and we had to stand within these kind of requirements. And actually recently we started also collaborating with either companies. So security is only getting better and better and the compliance needs to follow.

And yeah, I think that's ⁓ critical when you're dealing with companies IPs, drawings, models and so on. like, should be like in the highest priority. Michael Finocchiaro: ⁓ I don't have anything to add. Adar: No, I add to only to ORSAS that security is becoming a bigger challenge as more people have access to these tools and the attacks can be more sophisticated, right?

Because these tools are basically integrated with many things and they have external files and skills and stuff like that. And it becomes way, way more comfortable to share information. with them because it adds so much value. of course you need to take security very seriously, especially in our world of engineering and manufacturing where there's a lot of proprietary information, right?

There's a lot of files of customers that you need to deal with. And sometimes these files like you're signing non-destroysal agreements and... These files are not, you don't want them to get leaked to, that can become a national security issue in some cases. So you need to take security very seriously and put those measures in place.

Michael Finocchiaro: Yeah, I think that one of the issues we have is that we all started using MCP, but MCP had no traceability and almost no authentication, right? So we've talked about how AI was used in the development process and how it's in your stack. ⁓ We've now four years down the world from the inception of OpenAI and ChatGPT and all this stuff. Or Israel: So, that's my thought.

Michael Finocchiaro: Are you guys as skeptical as you were or as bullish as you were back in 2022? I how's your perception of AI changed as AI has been evolving over the last four years? Or Israel: and I'm sure the essential asset is the cost of our people. Thank you.

Adar: I think the way it evolved is quite phenomenal, right? I don't think anyone has expected this for us to be here. It looks like fiction. ⁓ If you took me a few years ago and you told me what we're capable of doing right now, the magnitude of what...

Michael Finocchiaro: Alice in Wonderland, right? Adar: what you can achieve with just some prompts, right? It's pretty amazing. I mean, I think everyone's, ⁓ even the most skeptical ones are getting convinced right now that there's something real here.

⁓ And I think the question that's shifting is, ⁓ for me at least, is not whether there is something. ⁓ real is like how we can be at the forefront and how we can utilize it the best way possible. ⁓ Because the fact that you have those capabilities and leveraging those capabilities is very different question. So that's where my mindset and I think many people are thinking this way.

⁓ The possibilities are so big. Michael Finocchiaro: There's a really good question in the chat from Sushil Kumar. And I think, Adar, it would be good for you to pick this up first, at least. ⁓ Sushil asks, as AI co-pilots start influencing design decisions and sourcing choices, how do you see ownership and protection of IP evolving, especially when designs are partially AI-generated and supplier-informed?

Adar: I'm sorry. Michael Finocchiaro: how do you see ownership and protection of IP ⁓ evolving, especially when designs are partially AI-generated and supplier-informed? Adar: Right, so yeah, that's a very interesting question. think.

As AI influences the design decisions, I would say, like, it's not only ⁓ unique to engineering, right? There's a creation is becoming something that ⁓ is a collaborative nature of AI and humans. Right? There's, I think that as we go further and further, there will be less and less creations that are solely ⁓ humans, right?

So we all have to get used to the fact that ⁓ our IP is basically something that we co-created. with AI ⁓ co-pilots and not necessarily like our own brains. ⁓ And we kind of think, the way I think of it is like we have a brain almost, right? That is living in our LLM and it already has ⁓ a lot of context about us and our preferences and what we like and the way we work with them.

So I think that... the way I look at it is like ⁓ the way us humans are still at least good at judgment, right? We're still very like I think that as we go judgment becomes a very critical factor and and the the IP here is basically like ⁓ the human gave the judgment on creating this thing together with AI, right? But eventually the IP is the humans because they created it and they signed off on it.

So that's way I see it because you can almost not distinct between something that's solely a human made, I don't think it's maybe even a relevant question anymore, to something that's co-created with AI. Even if it's just the human gave a prompt and the AI gave it an answer, the human made it their own. Michael Finocchiaro: I think my daughter was telling me that something like 60 or 70 % of the content on Spotify is all AI, which is just an astounding number, right? ⁓ Or do you have anything to add to that AI question or the question about ⁓ bullishness going forward?

Adar: is up. Or Israel: I just think that all of this motion of AI is, at least for the market that we're living in, it is a very strong boost. Because if you think about it, then you have the perspective and the experience from 30 years in this world and in this... ⁓ industrial manufacturing kind of sector.

It used to be very, very slow in the matter of adoption and in the matter of innovation in the side of the software, at least like ⁓ the adoption of companies and their ability to postpone like new software and development. And I think that because of AI and due to AI, ⁓ we see lots of opportunities and that companies started realizing that they need to move fast because everything moves fast. I like at least how I see it, I meet lots of companies that they have AI initiatives for the company led by the CEOs and lots of people they really want to make a change.

They want to change the way that they are working right now. So I think that it's not only the tactical way of how it is going to help and solve issues, it also shifts the whole industry towards a more receptive kind of vibe. And I think that's amazing for the industry. Michael Finocchiaro: I think so too.

⁓ I also like to ask at this point in the conversation, ⁓ what advice do you have to the younger people in the audience that are worried that AI is going to steal their job, right? Or Israel: ⁓ So, AI is phenomenal. It's only going to get better and better. And I think that the people that are using AI and know how to take the most out of it are the ones that are going to lead the next generation of positions.

at least when we're dealing with tasks that are manual and we automate them, obviously it makes shifts. like marking up red lines, I had like 30 mechanical engineers in my former ⁓ position. And some of the people, the majority of their jobs was to verify that the changes that were documented between two drones were like corrected before submitting it to the FDA. Now, obviously, that becomes like an obsolete because it's like you can't defeat a computer doing that, obviously.

And what people like are going to do is like the more interesting stuff like creation or like being able to do like physical tests or create more and develop more. So I think that people are going to make a transition into ⁓ like the the cherry on top kind of work. And I think that for junior engineers that are now looking to see how they're going to go and work in the next generation of positions. They need to take a look on the industry.

They need to understand where things are headed to. And they need to adjust. They have a smart young brain. And I'm sure that they're going to understand where are the trends and how to take leverage off of that.

Michael Finocchiaro: So do you have anything to add, Ada? Adar: I think the younger generations have a massive advantage actually. They didn't spend many years perfecting a certain skill and they actually ⁓ have somewhat of a clean slate in many ways to think in a way that the industry is rethinking itself. ⁓ Michael Finocchiaro: Hmm.

Adar: So I would say like if I were to in the beginning of my career, I would definitely take advantage of that and learn these tools and perfect how to use them because there's definitely like ⁓ there's a learning gap, right? Not everyone is operating the same way. And the second thing is develop a good judgment and taste because that's, I think, for the foreseeable future and not going away, still humans are controlling the output. So if you let an LLM...

and run 10 scenarios and you pick the right scenario and you tell them exactly what you need changing, that's not going away to understand the professionalism. So that's why the top engineers are just becoming better engineers and the top designers are becoming better designers because that thing, that taste, understanding of what's good and what's not is not going away. So you think of it like a chef in a restaurant, like they're not necessarily cooking, but they know what a good dish looks like and how to ⁓ change the dish to make it a little bit better.

⁓ So developing the taste in your profession is a very important thing. And I would look at it as an opportunity, not necessarily as something that you want to be scared of, but actually like something that you can be excited about. Michael Finocchiaro: It's a really unique answer, Adar. Thanks.

I usually get fundamentals, math, physics. That was a really cool way of looking at it. I appreciate that. Before we close out this section and we go on to digital maturity, I just have one last one for you, a kind of fun one.

Have we hit the open-air moment for engineering? Or is it six months away? Is it two years away? I when do you...

Because I mean, there was an open-air moment for AI, right? I mean, that was... There's a before and an after. When are we going to see that in engineering, you think?

Adar: ⁓ So when you say ⁓ open AI moment for engineering, what exactly do you refer to? Or Israel: So we're gonna stop. Michael Finocchiaro: Well, the way we used to work and where they work now, they're, you know, they're pivot, uh, moment when nothing is the same. I mean, after open AI, nothing is the same, right?

Everybody's life changed forever. So what are the engineers life going to change? When it fundamentally where the way we do engineering will be different. Uh, I don't think we've changed.

I don't think that's changed yet. I ask also because I was at prove it, which is a manufacturing conference and I saw them deploying entire factories. Adar: Great. Yeah.

Michael Finocchiaro: for looking at the sensors in less than three minutes. And that's an open-eyed moment. I mean, it used to take six months to a year with all these sensors and half the sensors. Of course, you couldn't read them anyway.

Now they can deploy that like that and have the entire visibility from top to the bottom of the factory instantaneously. That's an open-eyed moment. So when for engineering will we see that pivot change, ⁓ was the pivot point? Adar: Yeah.

Right. So, so when I think about a major fundamental shift like the internet, like it took a while since the internet started to with all the ways for people to change how they work, how they do things like new companies started, like that, that shift took years. I think humans even got to use and understand the, the, the, Michael Finocchiaro: The implications. Adar: the implications and how, you can do with it.

I think that, so for certain use cases, takes more time, like engineering, construction, like things in the physical spaces, just companies and. and use cases and tools can be built around even the existing capability. Even if AI want to get better, it will get better, I think, in creating CAD design. We see some start of it with Opus 4.

7. Suddenly, you can get pretty impressive. Michael Finocchiaro: Yeah. Adar: things, but of course there's long way to go.

even without that, the model shifting and getting better, think that just the adoption and people starting to understand how to do things, it's going to take a while even when the technology is evolving in such a rapid pace. Michael Finocchiaro: Thanks, Adar. Or maybe same kind of question, but maybe is it going to be CAD as code that's going to change it? Once we can get the 3D geometry as a textual thing that an LM can read, is that the big change that's going to shift it?

Or do you see it somewhere else? Or Israel: So it was a first for me. I it was going change. I think that the trend already started and like the difference is in the adoption of open AI is like a B2C kind of product, right?

Like you could use it like on your own computer, trying to understand how it works. In our world, it is B2B world and you need to go through the company in order to get into the real problems in the real data. And it moves ⁓ slower than the open AIs, but I think that we are already there. I think that with some of the companies that we are working with, you could already see the trend, the motion, and the impact, and you could understand how scalable this could be.

And if to be like, I'm an optimistic, so this is why I'm a founder of a company. I wish to say like... 2026, maybe like the beginning of 2027 would be like the time that we are going to see like a significant use cases of a huge impact. Michael Finocchiaro: Awesome.

So ⁓ the last section I like to talk about is ⁓ concerning digital maturity and of the customers that are using your software. ⁓ I think there's a global feeling around engineers that the big software companies, the big dominates on the market are not necessarily delivering the goods on AI. They're a bit slow on adoption. so customers are bit frustrated.

And I think when you look at digital maturity, think of ⁓ companies being on a scale of one to five, right? One, they're still using Excel and email for almost everything. Five would be fully agentic, autonomous, adaptive digital twins. Nobody's at five.

Maybe some pieces of SpaceX or Tesla maybe, but even then the rest of the company is definitely not at five. ⁓ How do you assess your companies when you start talking to your customers? Do you feel that they're closer to one, closer to three? mean, where would you put them before they start using Bananaz or Jiga?

Or Israel: I would say somewhere between two to three. That's ⁓ like the type of companies that we usually try to find as design partners. I think that this is also an indicator of the company that helps to get the software and get the value from the software. When a company has one, it is very difficult to go and sell them the product because they can't even...

Michael Finocchiaro: That high. Or Israel: ⁓ digest what they have in their hands. So that's like the type of companies that we are looking for and five that's for companies that are startups or as you said, companies that are relatively small. Usually it's for us, ⁓ it's not exactly the ICP right now.

We are targeting companies that are large. this is the type of the optimum, getting them at least two or three and being a big company. or at least a part of a big company. Michael Finocchiaro: Have a bad four, Jiga.

Adar: Yeah, I would say same, like 2 to 3 is a good sweet spot, so same answer. Michael Finocchiaro: Okay. ⁓ And then the second part is my theory is that when the customer, when customer start using a really awesome software like Jigar Bananaz, that that moves the needle, right? That it moves the bar towards the right.

And perhaps ⁓ the parts of the organization that are not using Jigar Bananaz see that, my God, if I was using AI correctly, if I had broke the barriers between manufacturing and supply chain and engineering and manufacturing, I'd get this incredible boost in productivity time to market profitability, quality and so forth. Are you seeing that already? Are you seeing that when people adopt JIGA and Bananaz that there's this on the part of the management? Like, my God, if we just did things correctly and we caught the 21st century and our infrastructure, our architecture, our thinking.

we'd these enormous gains. Or Israel: Definitely. think that this is one of the ways that we make ourselves. It's not only to provide the value, but also help the management ⁓ to appreciate that and see that.

It's two separate ways. One is to calculate the return on investment and make sure that they see that it is significant, but also understand how impactful it is for their businesses. Sometimes these kind of things, especially in engineering, it takes time because you can see immediate value. But in order to actually appreciate that, you need to go and create a full engineering change order process or a quarterly process.

You need to understand how it is affecting the different departments and the whole ⁓ development cycle. Eventually, the development cycle takes time. There is manufacturing. There are supplies that needs to go and ⁓ like...

⁓ build either in the manufacturing process of the parts or the assemblies. ⁓ But eventually it's a loop. Eventually you get into a process that you understand and you could quantify how much time you were able to save in the time to market. Michael Finocchiaro: Very cool.

And similar idea when people are starting to use Jiga. Adar: ⁓ So for us, I think about this from the perspective of like habits in general, like building habits. ⁓ And it's very hard to change, right? Just in general, changing how you work and how you do things is hard.

And there are ways to build habits that if you introduce something that is ⁓ very disruptive or changes how you do things in very strong way that you will just go back to your defaults. Even if it's like, even if all the logical explanation is there, ⁓ sometimes you still go back to default just because it's the default. It's hard to drive change. It's hard to start working out.

It's hard to these things. Michael Finocchiaro: Sort of hard-coded in your brain already, Adar: Yes, exactly. So as you understand how our brains work and our brains build habits in general, ⁓ so we try to work according to that. And that means, for example, starting gradually, doing things, creating small wings before creating the larger wings, for example, or making it as easy and seamless as possible.

attaching where you can to how things are used to be and just not having necessarily the ambition to radically change how people do things but kind of blend into how they do things where it's possible and deliver value there. So like this type of stuff, ⁓ they create a much easier way to adoption and then people start as they get into that habit loop and... and these doses of dopamine from weaning, example, they would expand their usage and they would just beat from there. Michael Finocchiaro: That's awesome because I think that it's important to find those ways of accelerating innovation and showing that you don't have to buy from the big three in order to get the big gains.

can get the big gains to the startups. I've looked at the market and I've talked to lots of founders and of the 600 startups I've talked to, I can show that they're implanted in 1,500 different companies, which is pretty astounding. Adoption rate that's coming towards the end of the call. I don't know if you guys have any closing thoughts about about Bananaz about AI about Giga Before before we say goodbye Or Israel: Yeah.

Michael Finocchiaro: you I thought it was a great discussion. appreciate the candidate answers and ⁓ it's exciting to see what you guys are doing and it's ⁓ so courageous and it's not easy right now. I hope to see you guys hopefully at a threaded conference in the future. I probably will have one in Munich.

That's not too far away, right? ⁓ Sometime in the fall and maybe I'm booking it, maybe Denver. Or Israel: And it's a draft of this. And it's draft of the way that we're going say that.

Michael Finocchiaro: Something maybe in the United States towards the of the year. ⁓ So once again, thank you very much for my guests. Thank you to AWS, our sponsor. You'll see a link in the comment.

Please click on it and download the white paper when you get a chance. And we'll talk to you the next time. think the next time I'm going to have, we're going to talk about virtual reality. I'll have gravity sketch and campfire.

So it's going be really interesting to see how virtual reality is coming into engineering and changing the way we do stuff. So. Or Israel: So that's it. I hope you enjoyed that presentation.

I'll see you tomorrow. Bye. Thank you. Michael Finocchiaro: Once again, thank you very much, Or and Ada.

Or Israel: Thank you very much. Thank you very much. And if somebody is coming to either the user group of PTC in Vegas next week, that's like we have a booth there. So come and jump, say hi.

think it's very good way to start both making collaborations and discussions about AI and about how you utilize that. the organizations will be able also to see Adar: Thank you very much. Michael Finocchiaro: ⁓ yeah, was going to say. Okay.

Or Israel: our product and we are running like live demonstrations there. Same goes about hardware, FYI conference in San Francisco, which is held in second week of May or IMTS in ⁓ Chicago in September. Michael Finocchiaro: ⁓ those guys, yeah. It's Chicago, Chicago in September.

Ed, where can we see and Jiga Adar: ⁓ There is a Boston Robotics Conference coming up. It will be there. then, yeah, potentially IMTS as well. We'll see.

Michael Finocchiaro: Nice. Okay, well great. I look forward to meeting you guys in person one of these days. Take care, be safe.

And for the audience, thank you for joining and we'll talk to you next week. Thank you. Or Israel: Thank you very much.