Creating a Visual AI Interior Design Solution with Alon Gilady of Renovai

DISCLAIMER: Below is an AI generated transcript. There could be a few typos but it should be at least 90% accurate. Watch video or listen to the podcast for the full experience!

Alon Gilady 0:00
And that’s basically the reason that we’ve become a is understanding design in a fundamental way. What we are doing we are trying to understand design as if we are taking the AI machine into a design school

Alexander Ferguson 0:20
Welcome to UpTech Report, this is our applied tech series. UpTech Report is sponsored by TeraLeap. Learn how to leverage the power of video at teraleap.io. Today, I’m excited to be joined by my guest Alon Gilady, who is based in Tel Aviv. He’s the CEO and co founder at Renovai. Good to have you on, Alon.

Alon Gilady 0:38
Hey Alex, pleasure to meet you.

Alexander Ferguson 0:42
Renovai is a virtual AI based home decor designer. So you guys integrates this solution into retailers who are focused on selling furniture and homeware. Fast name. I’m intrigued to hear more about the technology. But this the problem? What was the problem that you initially saw in this space and wanted to solve?

Alon Gilady 1:02
Yeah, it’s a great question. So there are two main problems that we saw. The first one was the end consumer, we understood that most of the people that try to buy furniture, try to decorate their home, and not using designers and spending a lot of time in online research because that too many options to choose from. And they have a significant lack of confidence in the capability to design many items together in finding the right items. And for retailers, we understood that retailers main problem or main goal is trying to do upsells and cross sells and trying to create an experience that is more holistic, more design oriented, compared to what they currently have. And I think our solution is able to support these exact problems.

Alexander Ferguson 2:01
Well, I’m intrigued that we will get more into the technology and how that works. But let’s take a second just go back in time. You and your co founders to CO for other co founders Correct? Yeah. 2018 I see listed December 2018, on your LinkedIn profile is when it started. But what was that journey like before this? Where were you? How did you get into into technology and building a design, AI design tool?

Alon Gilady 2:25
It’s really interesting. So the story starts, when myself and non children with a co founder with me. And another friend of ours named Andre started a 3d studio. So we have created visualization and 3d, a apartment building for architects and interior designers. And we ran a small studio. And from that experience, we really understood the gap in agency, we understood how expensive their services are. And we understood that most of the people do not go with your designers. And from that experience, we thought there is no way that using vision and AI, that machine cannot learn the principles of how to design how to make items together. And that’s what, you know, motivate us to explore and figure out if we could do that. And the

Alexander Ferguson 3:25
idea was this that you were doing this 3d designs to Yeah,

Alon Gilady 3:28
I would say between 16 to 18, something within that range. And in when we thought about the idea, my partner Luke German thought about the idea. And I talked with my brother about verifiability my older brother, he is a MA in physics and a lot of experience in the field. So I explained him the idea and he decided that it’s a good vendor to explore. So originally you Val join us and then we get to know of their of their is a colleague of his in a in a company that they used to work with a of now has a PhD in physics and one of the top experts in computer vision and new network in Israel. So that’s how the four of us came together. Unfortunately, Valley is not actively a part of the team, but that’s how we we got together.

Alexander Ferguson 4:25
Yeah, that guy which kind of connected to an unknown and kind of like the tech side, you’re developing overall product and leading as CEO and then alone, the other alone, alone. Is customer kind of oriented and operations helping us together. Guys have divided it up. Did you already like from the beginning? No. Okay, you’re going to take care of that. You’re going to take care of that. Like how did when in 2018 after the first couple years 2016 2017 like this could be a solution and these pieces come together? Was it just you meet together and say alright, this December, let’s launch

Alon Gilady 5:00
No, definitely not. And to be honest, between the two of us, we are still having traveled to divide the tasks. So we are doing a lot of things together, it was really obvious that I’m going to lead the company because some of the human skills that I have in the fact that I have some experience in leading teams, so that was a big factor. And Alon is one of the most creative people that I know, is always thinking about you out of the box ways to innovate. We knew that he’s going to be more of a taking that for but they usually we have we are we have done everything together, basically. And have no obviously was on the tech side, it was very, very obvious what he’s gonna do.

Alexander Ferguson 5:52
And yeah, actually building building the solution. So yeah, it’s like 2018 is when you begin to build it, how long did it take to actually come out with the product? What was that timing like?

Alon Gilady 6:03
So in 18, we’ve started to do that part time, I would say, A, trying to create an MVP, we understood that we have to have some kind of technology, it’s not going to be enough to have a deck, at least for us, you know, young founders, a first timers, we knew that we had to be a walking, walking product. So yeah, we’ve built the product. And by the end of 18, we presented to a few micro VCs and angels. But I think the game changer for us was being selected into Alchemist accelerator in the US, which I think was one of the best moves that we have done. Both in terms of think more knowledge, getting the mentors that they have getting the validation for the internal team and for a potential interest investors. And getting top advisors that could really guide us through the process. That was B, that was really, really prices.

Alexander Ferguson 7:09
What year was that? That you went to the accelerator?

Alon Gilady 7:13
So it was end of 18? Beginning of 90?

Alexander Ferguson 7:16
Right? Right at the beginning piece getting that getting that rolling? Yeah. Yeah. When you were looking at in the space, and still today, I mean, he’s what’s the competition? Like? Is are you like seeing and you’re like adapting? Or is there? Is this still new? What does that look like?

Alon Gilady 7:31
So definitely, there’s a lot of competition in the area of what we do. A lot of people mistakenly compare us to, I would call it Do It Yourself application that provide 3d visualization tools, there are a lot of companies that create, you know, either augmented reality tools or 3d Aiko tools. But we are not there. What’s unique about what we do is to understand design in a very fundamental way I can describe the technology later on. But I think what’s unique about us is understanding design in a fundamental way as if our machine was an interior designer, and combine it to the visualization part, combine it to either 3d experiences or editor’s experience or even augmented reality. But we always try to sell what’s unique, our core IP. And I think that’s why top brands around the world are really excited to work with us. I think we bring something new to the table that no one else thought about the problem the same way.

Alexander Ferguson 8:40
But let’s let’s we might come back again, to hear more of the journey because I know there’s a lot of peace I’m sure happened since you went to the accelerator in 2019. So now, but let’s talk about the technology the product for a moment, because it’s there is a big push for more online e commerce, yay COVID pushes technology forward. But your approach to this is both using a visualization of the product, look, look at this, this product is proud of this product, they could all go together in a home environment. But it’s also it’s both showing it but knowing the right items that need to go together and the tagging of set items like tell me like what’s the workflow that you’re you’re providing?

Alon Gilady 9:21
Yeah, I’ll try to describe it. But we have we have three main layers of technology. The first part is understanding what inventory we have, we are working with retailers and we need to understand what we have in hand.

Alexander Ferguson 9:36
While you’re describing this, can you can you give me like an example what are your customers so that we can have like like, is it a furniture store or something so we can Yeah,

Alon Gilady 9:43
furniture store like for example may.com which is one of the biggest retailers in the UK. A so they provide us the inventory that they have and our computer vision is able to tag the most important attributes for each item based on The image in color the checks the material, the texture, style of the item recording technology that we have created, is the first step in order to understand what we have in hand. The second step?

Alexander Ferguson 10:12
Yeah, let’s just take a step back just for a moment there because we’ll get more steps. But style, having a machine understand style, you say that’s proprietary that others aren’t doing that, how have you approached teaching the machine style?

Alon Gilady 10:29
It’s a very challenging process. And we had a lot of, you know, rounds in order to understand what exactly do we need to do. And basically, we were able to break down for example, for each category, we could call it sofa, we have breaking down sofas into archetypes, basically understand style of a few combinations of features into one group, we’ve broken down each category to dozens, I would say, more than 100 Different archetypes. And then our machine is able for a new item that comes in, it is able to understand which of the archetypes is similar to the new item that comes along, understand the main features inside of it, and so on and so on a so a very, very deep understanding of arguably

Alexander Ferguson 11:26
is his style like, this is a modern sofa, or like what if one of one of the tags was like okay,

Alon Gilady 11:33
much more detail. Obviously, we’ve tried initially to understand this is a modern sofa. But then we saw that the variance between different sofas within the same style is huge. And what we’ve done is creating archetypes basically saying, for example, this is a tuxedo sofa, it has specific back, it has specific computer knowledge. It has specific type of an arm, and so on and so on. And so we have a lot of attributes specific legs with a lot of attributes within that, in order to understand the style of the item on top of you know, understanding different features.

Alexander Ferguson 12:09
Are you a designer like bread? Like you guys just naturally that way? Did you have to bring designers on team and how have you approached this?

Alon Gilady 12:17
That’s a great question. No, I’m, I learned economics and philosophy. So this is not my background. But yeah, we have a team of interior designers and architects in what we do. And we bring a lot of, you know, methodology and knowledge from real people.

Alexander Ferguson 12:36
So let’s come back again to the stages so the first stage is a furniture store has a catalog and they already had the digital images, probably flat 2d Not not too leading. And then your machine learning goes through it looks at it and figures Okay, color, style, size, shape, style, being a big piece, you’ve already created that custom piece, the probably the other ones already exists, you were able to find algorithms for that and assign the tag so it knows what those items are what’s the second stage

Alon Gilady 13:02
second stage is to understand the shopper. So basically, what we do are interesting kind of a pixel in check all the previous actions, the customer items are added to the wish list to the car purchase before interacting with we take all of that under consideration. And because we have a lot of attributes for the entire inventory, we can truly create a unique profile for each individual shopper, understand the practices in every in every category, understand the preferences in in in styles in colors, and making a lot of educated guesses about what we should offer.

Alexander Ferguson 13:42
Okay, so that’s the second stage.

Alon Gilady 13:44
And the third stage, I think that’s the most difficult and interesting part of what we do. And that’s basically the reason that we’ve started a company A is understanding design in a fundamental way, what we are doing we are trying to understand design as if we are taking the AI machine into a design school what we have done, we have divided the entire design, interior design world into styles. And each side we have collected 1000s of different web design images from our catalog, the topic designs, and then how our computer vision was able to take all the items inside of all the relevant items. And using a knowledge graph, we’re able to understand the correlation between the items and of parliamentary materials and shapes and texture and basically with that capability creating design knowledge. And once we have that design knowledge we know the shopper and we know the inventory. We basically have everything that we need in order to create customized conditions.

Alexander Ferguson 14:49
And then in this visualization as it puts it together are you then literally pulling those images put them together and say alright, this is a suggested living room of with a with a chair and a rock gonna plant and or whatever.

Alon Gilady 15:03
Yeah, so we have a variety of different tools, we try to be a one stop shop solution, or retailers. So we have a tool that helps to find one item inside category pages, we have a tool that if you’re on a product page, using our tool, we are able to say, look, we’ve had so far by that rug that under that pillow, and so on, and variety of ways that we’re going to present. The first one, if the vendor if the retailers that are working with us have 3d content, wherever to create photorealistic rooms, inspirational settings, I would call that we call present all the recommended items inside a beautiful roll room. A. And for retailers that do not have 3d, we are offering a very, very interactive moodboard we can call it taking pieces of truth to the flat images into a billboard. And we do that interactively. Customer obviously are able to swap items was tentative and play with it until they’re getting the perfect results.

Alexander Ferguson 16:09
So really an end to end solution is what you’re you’re providing for for a retailer.

Alon Gilady 16:15
Exactly. And we try and not only that, because we understand that customer journey does not stop when they leave the site. Another great solution that we have is a remarketing solution. So check customer that, for example, edit a few items, items, which does not convert, we take these items, inject them into one scene or a mood board compatible with a complementary items, and then creating a landing page for each individual shopper and creating an image for the retailers to use in a remarketing campaign. So imagine that instead of having a sofa, a country with different websites, you’re going to have full design solution that is perfectly created for you to make the experience much more design oriented, and much more personal compared to what

Alexander Ferguson 17:10
what kind of fascinates me when I when all these different discussions that I have with different tech leaders is like, as a consumer, we make expectations when it comes to businesses that they actually already have this technology. When it comes to recommendations like on Netflix we have such we assume the next show we see. That’s what I’m going to like because they should know. And what’s funny is, in many cases, many businesses and different entities don’t have that yet. And that’s what you’re trying to design is is that experience.

Alon Gilady 17:36
Exactly. And one more thing, Netflix references really, really, because a lot of recommendation engines that rely on I would say big data and put you into segments. So for the sake of example, it understand that I’m looking at so far and another 1000 people after they saw the so far they might purchase. But we take it we are viewing that problem differently. I think that regards design, I think it’s more appropriate to understand design is if interior designers are trying to create and create a truly personalized experience. So we’re not putting people into segments we try to understand, okay, this is your personality. This is what you like, this is your budget. A these are the design elements and design principles that we already know, prior to any director. And based on these two, we can create truly recommendation recommended items for you. So I think that’s what’s unique about us. And that’s what’s different between us and our recommendation engines.

Alexander Ferguson 18:46
Do you did you say earlier on for that second stage that the customers are just filling out like a survey form for you to collect this information?

Alon Gilady 18:54
So it’s a good question. So, we have a few tools that they can have serving like a style quiz and based on that we are truly understand their specific needs. But for our main product, which is complete look inside product pages, we are just checking or the previous action they have done. So understand if you looked at a sofa and purchase a rug in an edit another item to the wish list. We take all the different attributes and trying to understand as much as possible from them in order to create a profit for you. A obviously as long as you are keep a engaging with a specific website, we can check more and more information to be more more and more precise form for your needs.

Alexander Ferguson 19:43
That is that the recommendation becomes smarter and the more you use it, and it knows what you’d like. They’ll give professionally AI professionally designed solutions into recommendation. So let’s come back to again your journey from Moment. So 2019 you go through the accelerator you go forward was has it been smooth? All the ways have been like, oh, yeah, we get it like all the retailers and furniture stores are like I want this, what do I sign?

Alon Gilady 20:14
In terms of clients, it’s still a challenge, I can say, initially, I would say it’s up to the founders to try to sell and I think this is number one advice that I can I can bring to the table is being stubborn, going into talking to retailers understand what they need, and to be agile with what you offer. And to tweak it to provide as much value as possible, I can tell you a short story. Initially, our number one product was a product called future design, it was the main goal for for the product is for customers that are moving to the apartment or trying to design a full space, doing a full style quiz and provide a full solution. I think it is a great solution. But we got to understand once we talked to a lot of retailers that this is a great solution that they might pay and experience. But it does not move the needle because the amount of people that are coming to the conference. And in fact, when we design a food space is extremely low, I would say below 1%. And then we were agile enough to think about, okay, how can we use our technology in a better position that can truly create huge amount of value for customers. And that’s how we thought about completely look inside product pages, because we understood that this is the position this is the pages that have the most significant amount of traffic into. And if we can move the needle and integrate an uplift off of revenue, even 5%. It’s tremendous success story. And I think that’s, that’s a big, big part of our success up up and below that you are agile enough to tweak our solutions and understand what the customer needs.

Alexander Ferguson 22:23
That is a powerful example of listening to the customer and not just building something and say who wants to buy it? For what what, what real is? Is that a been a recent one has it been like a year ago? How did that kind of come into play?

Alon Gilady 22:38
Took a while I would say that we truly chuck a started to sell in January this year, it has been a process learning curve, we kept developing the product until we are in a position to really sell it. So it has been a process in the end of 2020 or mid of 2020. We have ran free pilots with major retailers, they were all very successful, we have converted all of them into paying customers, or I would say paying partners, because these are our design partners. That’s why I like to mention me. And more importantly, one of our biggest customer named tempered web, sir, is one of the biggest retailers in Australia. Like the product in light of vision so much they are, I would say one of our significant investors in the company as well. So really a success story. And yeah, I could not be you know, it’s a blast to have these types of investors in the company.

Alexander Ferguson 23:43
That’s a great testament to the to the product when your own customers invest in your product because like we want you to succeed because we like it. Yeah. Though it’s an interesting combo is then there other customers will be other customers, potentially competitors could use your product, but it kind of moves the whole needle for everyone. Where you guys are now I mean, you’ve kind of honed in on the complete solution for retailer from ingesting the database of all their assets and then figure out their customers and then providing solution. I get excited about technology. But do you see how far in the future we’re gonna be like AR like augmented reality we have pulled in the assets. I mean, IKEA is doing a lot of apps right now and there. You can pull it in there anything you see that on the roadmap.

Alon Gilady 24:36
It could be again, I think that what’s unique about what we do currently is the recommendation parts when the reality and average solutions are getting more and more of a commodity so we can either invest in projects and to enhance that into our solutions. A later down the road But we want to double down on the capability to recommend items. A basically understanding design better and better, a maybe to expand to other verticals in the future. A medicals within the lifestyle. Interesting, I can say fashion, accessories, a jewelry, all the things that are related to visualization, I would say combining colors, materials and checks. These are the things that we weave a few tweaks in, in keep learning in training our AI, we can be very, very much we can provide a very, very significant value in these domains to

Alexander Ferguson 25:49
that makes sense with the potential expansion into other similar indices and markets. When you talk about like just that solution of like the mood board, it makes me think of Pinterest and things that what consumers are expecting to experience but in an in a retail space, you give them that same play and experience.

Alon Gilady 26:13
Yeah, definitely. I think that woodwinds are very customers are very familiar with a, they do expect to have that. And I think it’s absolution, I might have less of a wow factor compared to the mentality. But sometimes they because that was simple. And because people are really, really familiar with the concept can Jive sales in a more significant way. And we can see that sometimes with our existing products. When we compare what we call complete the look Woodwards and shop the room, which is basically the same with 3d. I cannot say that one is winning in a Justic way compared to the other, even though that, you know, the photorealistic room is more impressive.

Alexander Ferguson 27:05
It’s I love visual things, and I love futuristic things. So I always like to think that the the latest coolest futuristic technology will be the one that sells the most. But we know started like the bell curve of adoption. Yeah, people were like, well, this is the majority of people like I don’t this is show me something familiar. Familiar things. So

Alon Gilady 27:28
yeah, yeah, definitely. And sometimes, you know, the fact that you can play with a lot of functions, it gets complicated for most people, and they say it’s, it becomes more of a game compared to actually trying to figure out if these are the things that you need to try to figure out if they match each other. And sometimes, you know, simple is better a, except that I do believe that in the future, more and more retailers will have 3d content. And they will double down in the investment in these types of expenses. And we definitely need to be in the game of 3d and AR as well. So I’m not neglecting that in the long run.

Alexander Ferguson 28:16
Yeah, it’s just knowing where you need to focus now what were actually drives the sales for. As you said, it’s like this adoption curve, I feel like pre COVID Maybe they may not have been as interested but now you feel that furniture stores and home decor is this is definite must

Alon Gilady 28:35
definitely and more and more stores, you know, brick and mortar stores are moving to the online sphere. And they have to have an ecommerce because, you know, people need to buy online. And it’s also an education curve for for the consumers. You know, in percentage, the agency is moving online, it’s represent now the online. Ecommerce is representing around 25% of the market right now and it keeps growing. So the opportunity is definitely getting bigger and bigger and more and more stores are coming. It’s not only more people that are purchasing online, more and more stores are emerging into the E commerce.

Alexander Ferguson 29:23
You’re you guys are based in Tel Aviv. How big is the team today?

Alon Gilady 29:28
We are 17 people at the moment. A basic views a we have a sales a manager in the US and but the rest of it most of the team is it is well we have a few developers in the Ukraine as well.

Alexander Ferguson 29:47
Got it for you. When you look at this this industry, the space and what you’re building the product itself. What are you most excited about? What are you What are you looking forward towards

Alon Gilady 30:00
terms of potential of business or whether you fit,

Alexander Ferguson 30:05
I kind of, I’m leaving it somewhat open because I treat what you’re like what’s in your mind day to day like I’m very excited about where where we’re headed on maybe it’s part of your roadmap or what you’re building or where this is going.

Alon Gilady 30:18
I would like to be in a position, I think one of our major goals, I would say, is to automate the onboarding process in become a plug and play solution into the industry. It is a difficult task, because how complicated our solution is, and sometimes the visual aspect of that, A, but I’m getting excited if we’re going to be in a position that are going to be a plug and play. And we can, we can facilitate and work with 1000s of brands, instead of you know, dozens being available in the big, you know, big online platforms as a plug and play integration. I think this is one of the major goal for company A, this will allow us to really scale up.

Alexander Ferguson 31:09
You see them becoming more of a product led company versus sales lead. So the muka just Oh, yeah, let me make the run of AI plugin, and it goes to our database.

Alon Gilady 31:19
Exactly. Exactly.

Alexander Ferguson 31:21
Got it. Okay. They kind of closing here of where you see the e Commerce Industry going at any if you were to make a prediction, a technology prediction of where we’re headed, what can we see in like 234 years from now? What is it going to look like?

Alon Gilady 31:42
I think that too. Let’s say in their furniture, you see that to it, ways that I can I think that the image is going to go. First one is obviously more and more, as you mentioned, visualization tools and augmented reality tools that are getting better and better or virtual reality tools. In the end of the day, you might be able to see the entire, you know, catalog in your house in a very, very high quality photorealistic way. I think the interest is going into that direction. But I would say in the E commerce here, I think that website in the long run, will we look different? I think I think the so the user journey is going to be really personalized, maybe you know, pages are not going to look the same as they are now I’m talking a lot with Apple webs are how they visit product page to look in any believe that, for example, product page experience is completely broken. And it should be something that is much more interactive, much more personalized. So I I do think that the market is going to go into much more personalized experience in maybe what websites are not going to look the same as they are currently.

Alexander Ferguson 33:09
It’s a future of personalization, what everyone will expect. All right, show me exactly what I want. I don’t care what other people what other people see it show me show me what’s important to me. Yeah, definitely. So two elements of personalization, but also the the augmented reality or some people call the meta Metaverse Metaverse, exactly. company that has renamed themselves but whatever. Yeah. The the I’ll be curious when they would just put on a VR headset and Sara, let me go shopping from from home and then drop things in. We’re probably a little ways away from that. Yeah, not too far.

Alon Gilady 33:48
I think the main challenge of the industry is actually getting the assets, the 3d models of the tire inventory. Big challenge is even for big retailers. It’s a big, big challenge in for the smaller ones, you know, and they’re far away from getting the inventory. It’s really so once once startup company is going to solve that, I think the gap, it will be entirely vitual. And photorealistic is going to be much shorter.

Alexander Ferguson 34:20
Well, there is a company and I actually interviewed and I’m off the top my head that is doing that exact thing 2d to 3d. And they’re doing really cool things may actually probably connect with you. We’ll do that after the interview. But yeah, what’s amazing is everyone is doing like, like, we layer on top of each other. Each technology seems to support the progression. And it’s like the more focused you are, the more you can help help someone on that particular project. Totally agree. Thank you so much of them for sharing the journey that you’re on the technology and the future of home decor design with AI solution. So for those that want to learn more, you can head over to Renovai. It is RENOVAI. So Renovai.com and be able to explore more book a demo. Thanks again. It was great to have you on.

Alon Gilady 35:14
Thank you, Alex, thank you so much.

Alexander Ferguson 35:17
And we’ll see you all on the next episode of UpTech Report. Have you seen a company using AI machine learning or other technology to transform the way we live, work and do business? Go to UpTech report.com and let us know