Venturing into Fashion Tech

Build It Series: Achieving the Perfect Fashion Fit with AIMIRR's Pritesh Kanani

• Beyond Form • Episode 59

Share episode

From Mumbai to Microsoft to Body Sizing:
Today's episode feature's Pritesh Kanani, founder & CEO of AIMIRR. From his early days in Mumbai to steering technological advancements at Microsoft, we hear Pritesh's journey to developing virtual try-on solution. Initially kickstarted by his mother's wish to see his future wife in a wedding dress remotely, the solution came to life that is now deployed by fashion retailers online and offline around the world.

Patenting the AI-Powered Body Scanning Software:
Pritesh reveals the intricacies of the U.S. patenting process and how their patented sizing algorithm is set to redefine virtual try-on solutions with its ability to produce results with a 99% accuracy. Pritesh is so confident in fact that his solution can reduce online returns by 40% and increase average order values by 20%. Listen to this episode to hear what makes him so confident.

Learn more about AIMIRR: aimirr.com
Get in contact with Pritesh on LinkedIn: linkedin.com/in/priteshkanani

Send us a text

Support the show

--------
The show is recorded from Beyond Form, a fashion tech innovation platform. We build, invest, and educate fashion tech entrepreneurs and startups. We’d love to hear your feedback, so let us know if you’d like to hear a certain topic. Email us at podcast@beyondform.io. If you’re an entrepreneur or fashion tech startup looking for studio support, check out our website: beyondform.io

Peter Jeun Ho Tsang:

Hello, I'm Peter Jeun Ho Tsang, founder and CEO of Beyond Form, and welcome to the special series Build it, where I speak to founders and their clients as to how they're building fashion technologies. This series gives you a glimpse into topics such as personal struggles, technical challenges, working in fashion and more. On this episode, I speak to Pritesh Kanani, CEO of AIMIRR. Now, like many of our founders within Beyond Form, Pritesh also comes from a technical background, having studied computer science in India and then working for a decade at Microsoft, but he got the itch to start his own fashion tech company. Building has meant learning how to sew countless hours of scientific research and dealing with mountains of pape rwork by his us patents.

Pritesh Kanani:

Getting a patent is in in North America, especially in AI in 2024, is one of the hardest things.

Pritesh Kanani:

that, uh, you know something about it and I have prior experience working in the software industry. I have five to six prior patents granted in the space. I had the comfort of not being new into the space, but for someone who is looking to enter into patenting or designing a solution, I would first suggest to.

Peter Jeun Ho Tsang:

Let's get this conversation going with Pritesh on this episode of Venturing into Fashion Tech. How are you today Pritesh?

Pritesh Kanani:

Doing well, Peter. How about you?

Peter Jeun Ho Tsang:

I'm great. Thank you, looking forward to today's conversation. Fit and technology and AI is always a popular conversation on our podcast show. Obviously, you and I have been working now since spring 2024 and you've got lots of insights to share with us, what I'm sure our listeners will want to hear as well. I just want to set the scene for today's conversation In terms of AMIRR.

Peter Jeun Ho Tsang:

One of the elements is hardware, and if we're talking about hardware, we're talking about magic mirrors within the retail environment. Magic mirrors are not necessarily new. For example, brands like Uniqlo, Amazon Fashion, Rebecca, Minkoff have been actually using such technologies and exploring them since late 2000s. Their latest versions include AR, digital fashion and more, but few integrate real fit technology into them with accuracy. Amy, this is where it comes in and working its magic by combining all of these different elements to up the ante in fashion retail, whether that's online and offline. We're going to talk about that a little bit later on and really dig in deep into how you're making that actually work.

Peter Jeun Ho Tsang:

In terms of the global virtual fitting room market size, it was valued at 4.79 billion US dollars in 2023, and it is projected to grow to 25.11 billion11 billion by 2032, with a compound annual growth rate of 20.3%, and that's according to Fortune Business Insights. So this is very indicative of where the space is heading. A lot of retailers, a lot of fashion brands, are looking into it. The fact that it is growing means that more e-commerce players are also looking into it as well, and we're going to discuss about some of your clients a little bit later on. I think that's super interesting because it's a huge number.

Peter Jeun Ho Tsang:

You know, 25 billion dollars roughly with you know, in less than 10 years, means that there's a definite space to be had there, and where you're based in seattle, usaitesh, is certainly demonstrating that there is an appetite for search technology, since North America is dominating the global market at the moment in terms of virtual fitting rooms, with 39.04% in 2023. So, as I mentioned there, usa, that's where you're currently based, but you don't come from the US. So I want to just kick off this conversation with just thinking a little bit deep about your background. You come from India. You've gone to America to follow the American dream, studying at Chicago Booth, then working at Microsoft as a computer engineer. Tell us that journey, tell us how did you become an entrepreneur, changing continents and and actually ending up creating your fashion tech solution?

Pritesh Kanani:

Yeah, that definitely is quite the interesting journey. So I grew up in India, in the city of Mumbai, and I studied engineering back there from a university called Indian Institute of Technology, which is one of the most competitive engineering schools in India. After that I came here for my continued master's in artificial intelligence and I studied image processing and digital signal processing, which is where I solved problems like medical imaging, cancer imaging and zebrafish tissues very initially, and I joined the industry to solve a wider range of problems With Microsoft. I did spend about almost a decade solving problems in video communications and the space that is very in-depth in augmented reality Realized my potential for developing software applications at scale, which is kind of what became the thing. That was big and the United States has a lot of big tech industry and verticals that are focused on growing in that area.

Pritesh Kanani:

Talking about entrepreneurship and the journey at University of Chicago Booth, after my engineering I started college and entrepreneurship is a big ecosystem here in the US which is very differentiated. Being an immigrant, it is something that was not ingrained in me and entrepreneurship was not the first thing that came to my mind. Something that was not ingrained in me and entrepreneurship was not the first thing that came to my mind, but developing an essence for learning sales, learning other business aspects and the curriculum. It definitely ingrains in you the spirit of going it out and starting a business. So this was a very revealing opportunity for me to come abroad, start from scratch and build your own idea or your baby, and take it forward.

Peter Jeun Ho Tsang:

So spending a decade at Microsoft? You're practically furniture by that point. Were you just itching to get out of that corporate environment to do your own thing?

Pritesh Kanani:

So to that question, the answer is two-faced yes and no. A job at Microsoft is generally very respectful in terms of being considered as one of the top most engineers in the world in that phase, but at the same time, something to build of your own has its own vibe, has its own value, and to me, I wanted to get out there and try to build an own brand or own reputation for myself, which is where the entire concept of entrepreneurship was appealing to me, and learning more functions outside of engineering is something that was net new to me and challenging at the same time. So the itch definitely got me, or caught me, to go ahead and start something, and in terms of your entrepreneur journey, what came first?

Peter Jeun Ho Tsang:

The itch to do something new, market research, trying to find problems to solve, or accumulation of the two things, moving country? You know, as I understand, you actually started the idea by actually looking into. You know what tailors were doing.

Pritesh Kanani:

Yeah, so for me, it was more about solving a very specific niche problem rather than calling or giving it a coining, it a term such as entrepreneurship or being in a corporate job. For me, it was a problem that I desperately wanted to resolve, so it was the timing and the environment. In terms of the timing, the timing of it was during COVID, where my grandmother wanted to see my fiance in our heirloom family wedding dress and I tried to use a lot of software. I was trying to collaborate, accumulate and was very curious about why me, spending a decade in technology, I'm not able to solve this in a matter of minutes or find a tool that can do that. So that was what me to go in and care really deeply about it.

Peter Jeun Ho Tsang:

It sounds like you are a technologist through and through, by training and by trade, and this is very obviously indicative of somebody that has gone through that experience, of not only education, but also your experience at microsoft as well. Fashion, though, is a completely different industry. How did you navigate, entering the fashion world something very unfamiliar with yourself and what have been the challenges to date?

Pritesh Kanani:

yes, I'll start with the last part. What were the challenges? So initially it was a pet project. It was, it was a playground project that I I started doing until I actually started meeting people to sell and pitch like, like this is something that is cool, do you want to go ahead and use it?

Pritesh Kanani:

The fashion industry specifically requires a lot of credibility in terms of having designed or having worked on the ground. So in an example, the first Every potential customer that I wanted to work with was a fashion designer and they did ask me like what are your credentials in fashion? What have you done? What have you achieved? And I was blindsided. I was really stunned at that point in time as to how to answer. I tried to pitch myself as a technologist, but I thought everything being trained in the enterprise world, in the SaaS world, it would be as easy to have that conversation. But I was wrong at that point in time. So the best way that I started working on that is, I did my own fashion boot camps.

Pritesh Kanani:

I learned sewing. I went out in, you know, outside in Chicago booth at the Magnificent Mile. I went into local bespoke shops. I spoke to seamstresses and understood how pattern making, how sewing works and educated myself for the first eight to nine months on what are the specifics of designing in the fashion industry and how to talk the language that fashion designers resonate with. I think it is really essential in this particular space to have both a good understanding of the fashion, intensive knowledge and being at the ground working in the fashion space.

Peter Jeun Ho Tsang:

I could just imagine you in that first meeting with a potential fashion client. You're just being asked, you know, what is your fashion experience? I just don't know how you would answer that. I can imagine you were, quite you know, a little bit distressed from such a question, because, yes, fashion can be notoriously difficult to get into if you are an outsider. I never knew. Actually, you went to sewing class, though. Are you a good sewer?

Pritesh Kanani:

I am not a great sewer. I mean, if my life depends on it, I will get the job done.

Peter Jeun Ho Tsang:

You can repair your clothing. You know how to sew a button onto a garment, basically, which is a lot more than a lot of people can do, shall we say? So let's get stuck into AMA itself, the technology that this episode is about, and just for our listeners, this is the first episode of a mini series looking into fit technologies and some of the clients that AMA is actually working with. So set the scene for us. Pritesh, what does your technology do? Yeah, so AI.

Pritesh Kanani:

Mirror is an abbreviation for AI Mirror. What it entails and what we do is we provide a packaged software trial room solution for retailers and fashion brands. The trial room solution consists of two specific software solutions. The first one is a body sizing and scanning solution, where the user goes in and uses a camera scanner to find the right size and shape that would fit to their body for a particular garment that they are particularly interested in. The second solution that we do offer is a size-based virtual try-on solution, where a user can come in and, on themselves, visualize any garment to their size, shape and to how it would fit or flow onto their body. We use different patented physics engines and techniques to give users the right visualization, as if it were really a mirror experience at this point in time.

Peter Jeun Ho Tsang:

We've been working with about 30, 40 customers in North America and also more internationally at this point in time, and we are available on both online e-commerce websites as an independent app, as well as in hardware stores and actually, if we allude you to that, obviously the latter part of the name is Mirror and, as I introduced in the introduction, magic Mirrors and the technology itself, and actually they're not necessarily that new in terms of fashion, tech or innovation and I'm not sure.

Peter Jeun Ho Tsang:

Actually the mirror is kind of like the best way to describe what Haley Mir does, because you are a suite of tools and, from a technology standpoint, you are not necessarily just hardware, you're not necessarily just software, but you're accumulation of all of these different little things to make fit technology come to life, which I think is really interesting. One of the unique elements to Unifology is your sizing algorithm that you now have two US patents for. So congratulations for that, because that is not easy to get in itself and I think many people that are listening who are trying to go through that process will want to know a little bit more how you actually managed to do that. Take us through how you came up with a piece of tech that is patentable and what did the legal process look like.

Pritesh Kanani:

Yeah, to start with that. I mean getting a patent is in North America, especially in AI in 2024, is one of the hardest things that someone can imagine. And I have prior experience working in the software industry. I have five to six prior patents granted in this space. I had the comfort of not being new into the space. But for someone who is looking to enter into patenting or designing a solution, I would first suggest to go in and address what are the barriers. Get a good lawyer, understand it's worth the amount, it's worth the money you spend about like 10 to 15 grand. But knowing the entire process, end to end, doing the prior art search, is well worth the time and efforts saved when you start a company.

Pritesh Kanani:

So for me, talking about body measurements and sizing, a lot of these solutions, especially specific to fashion, were new and even though I knew a lot about the software.

Pritesh Kanani:

For example, how do you specifically measure the contour of a waistline for a specific individual based on their gender, their body characteristics, their shape, their lengths?

Pritesh Kanani:

Many people are tall, many people have different body measurement indices and you need to really get down to that study and reverse into the geometry and the calculations. In that sense, the next aspect of understanding a patent is you need to not go too deep into the technical algorithm, but rather create a design that someone who would be an expert would be able to create, or build your product, and then just articulate it technically. A bunch of lawyers come in and help you out in terms of going through that process, and the last thing that I would say is have the patience to review each and every thing yourself, each and every description or detail of how your solution works. What are the scenarios that you are thinking, not now, but five years from now or 10 years from now and put that all together in terms of your idea or description. You don't have to be technical to write a patent. You just need to understand your solution. That's the bottom line in that case.

Peter Jeun Ho Tsang:

You make your science so easy, pritash, but actually you did a lot of work before you reached that stage. For example, you did a lot of market research. From my understanding, and that market research and the academic research I literally informed okay, what do you then put into that patent paperwork?

Pritesh Kanani:

Yes, so that is true. While I did have a lot of scientific knowledge in the software space, I made a conscious effort to go to conferences and work with fashion professionals. For example, I've worked with a computer scientist from University of Manchester and also the lead chair of IEEE, who sets the body standards and scanner associations. So an exercise that I was asked to do, even to be a member of such prestigious organizations, was to compare myself or compare my ability in terms of can I develop a scan and submit a technical solution that passes the minimum accuracy score, which was around a 97.9%? So my first submission was around like 80, 85% and I had to spend about six months iterating and making the solution better to be able to consider an expert in this field or technology.

Pritesh Kanani:

After that point you need to iterate and contribute, start contributing into the different segments and contribute to the society and the industry vertical. So it is a matter you need to put in the hard work. There is no exception for going into any field in general, but in particular for entering any new venture and for me it was passion. It is best to connect with a network of professionals who already have begun academic research or patentable research in this field and that would accelerate all the efforts that you take in this domain. I think that's super interesting.

Peter Jeun Ho Tsang:

You call yourself an expert there, tash. Of course you are an expert in fit fit technologies, but I think that is a sign of to file a us patent, you have to literally be an expert in what you're filing on to make it credible, to validate all of these different elements, even to go through the legal process in terms of then, you know you have. However, it's still incredibly difficult. You know we've had other founders on the podcast show previously that have talked about this issue. You know they've tried to tackle fit sizing, virtual try-in and so forth. You know, from different angles, from different perspectives, and fit within the industry is still one of those gray areas that everyone's trying to tackle but nobody's necessarily cracking as a code there. It's very difficult and there are just so many variables. What indications or indicators have you seen today in your MVP when you've been building that has led you to the right path of solving the issue of fit?

Pritesh Kanani:

So let me start with the fact that solving the issue of fit, while fundamental, is one of the hardest technical problems there are to solve. You need to first get body measurements right from most of the users, from a wide user base who own smartphones and laptops and desktops or no special purpose cameras, and then, at the same time, you need to find the right industry sizing where there is no standardization in today's day, like even the size charts or technical packs that retailers and brands use. You know, something medium in one store could be bracketed as a large in another, so that complexity of translating measurements into brackets and eventually looking at the sizing is a multi-pipeline project in itself. So that's the reason this research has been evolving over the past five to 10 years and we are still industry is still advancing towards getting the right solutions. From a perspective of the indicators like why did I, despite this amount of dense research in the space, why did I decide to move forward and also work on these solutions? It's because there was a specific niche and a need that I was able to solve with this particular technology.

Pritesh Kanani:

I take the example of plus size brands, who would also like to get their users to accurately measure themselves in a special BMI range or in a special class range.

Pritesh Kanani:

Although there are a lot of solutions out there, there are very limited solutions in that particular space that do the job right in terms of getting the right size One arm could measure separately from the other arm and that level of granularity, that level of 3D reconstruction Off the body, takes like years of work to perfect and you need to really be persistent in getting that right. And our fashion industry has plenty of use cases for each and every niche in order to move that Also. Another indicator was the financial investments. A lot of investments are being made into the fashion and sizing because of the high degree of returns to the fashion and sizing. Because of the high degree of returns, multiple fast fashion brands are losing about 20 to 30% of their revenue in this. So that was my leading indicator that this is a real problem to solve and some early successes in plus. Sizing brands, like large size brands and also made to measure, were my signs of like going in and solving this problem in those spaces.

Peter Jeun Ho Tsang:

I think, as our listeners have guessed now, you are somebody that takes research extremely seriously. You like to look at every detail, and when I was looking at your project with the Beyond Form team, you know when we're looking at if we were going to work together in a partnership or not. You know, one of the things that struck myself and the team was that actually, the amount of sheer volume of research you had done, the fact that you had validated so many points from a not just a commercial perspective, but also from a technical perspective as well, which is impressive to see, you know. Going back to the points about having two painters as well, I think this really validates the fact that fashion tech can be validated legally as well with those painters, and I think that's a really strong point for our category in our space, because quite often people take fashion tech as something frivolous, but actually, as you're describing here, no, it's not. It is patentable. It can be taken seriously from a legal perspective.

Peter Jeun Ho Tsang:

There's a lot of academic, scientific research going into some of the topics that we're talking as well, and whenever I speak to a potential startup or a fashion tech founder, I'm always asking them okay, what areas of your potential solution or product are proprietary to you. What is that ip? What is making unique, so that you're not just, for example, using somebody else's white label solution and just repackaging it and calling it your own fashion tech startup? And I think that is super important, as my takeaway anyway from our conversation is that what is needed to be done to really make sure that you're doing something credible there, and because you are credible, of course that translates into commercials as well. It means you are getting clients, big, small, all over the world. You know you've just come back from sourcing trade show as well, which I know that you've had many positive conversations as well, which I'm super happy to hear. Some of your clients do include Raza Baladana Shamshik, just to name a few. Can you share some of the success metrics today and how will they inform your business roadmap moving?

Pritesh Kanani:

forward? That's a pretty detailed question and I'll try to answer in terms of how these successes have translated in. So, in terms of the proprietary research that we initially start, my methodology of going into any particular research is to get to state-of-the-art first and collaborate until that point with the best researchers, the best scientists, and advance the industry incrementally from that point onwards. So in one example, we did have a scanning solution that was pretty good in terms of providing the measurements, but the incremental improvement, or the patentable improvement in that aspect was how do you visualize or wear that particularly sized garment onto your body? So the pattern that we did cover in this space is how do you interact with a garment? Or, if it's a dress, how do you calculate the flow of that dress, geometrically or volumetrically, after you get the right sizing, and how an end shopper perceives it. Such are the improvements that we make. We patent it and then we ensure that we are not infringing on someone else's work. And if we are using someone else's work, we absolutely 100% license that work and then build on top of that. So that's a fundamental approach in terms of making those improvements.

Pritesh Kanani:

Talking about customers, an example of ValorData was my first. We launched our first product together at TechCrunch Disrupt in terms of visualizing a Blue Raffle dress, which was one of their best sellers. What we did observe on that was you couldn't get past 20% returns and you couldn't get past 2% to 3% conversion rates, and in the industry, what we really wanted to achieve was to get to that 7%, 8% conversion rates. Our technology indeed, in some of the pilots that we did on Drop for Thanksgiving and other areas, we were able to sell 11% more and we were able to increase those conversion rates, bring it up to about 4%. Those were the early indicators of like those incremental improvements that our customers saw and then they put faith in us with regard to the technology in that space.

Peter Jeun Ho Tsang:

So it seems like these small clients to get you off your you know, your feet off the ground seems to be proving fruitful. Going back to my introduction, again, you know this seems like it's demonstrating that the space is growing. You know we talked about a 25 billion market by 2032. I just want to go back refer your wife protest, because I know she's been so instrumental in helping you to develop the technology itself. You kind of alluded to it earlier but she is pretty much you know every scan that I see there's your wife in there. Like how was you know that family man? How has she been so pivotal to you?

Pritesh Kanani:

so there are are plenty of detailed elements to it. So Garima, my wife she comes herself, comes from a textile background in India. Her family wholesales and distributes garments in India and she's truly been deeply passionate about the space. So she in her wildest dreams did not imagine that I'm ever going to present an innovation in fashion or work in this space for any amount of time. So that was the most exciting part. So she's been with me in terms of product testing, helping out marketing and doing things silently in this space.

Pritesh Kanani:

In entrepreneurship, when you're a founder, your silent co-founder is your partner. In this space, in entrepreneurship, when you're a founder, your silent co-founder is your partner. In this case, during stressful situations or during those stressful demos or overnight software solutions that I would present it, she used to ensure that my laundry is done, I'm well-fed and I'm ready for well-represented demo day and everything on the front looks good. So that's the reason I personally consider her visuals and her work lucky for the company and a lot of customers. It gives a personal touch to a lot of our individuals and the actual story of the beginnings that we've had.

Peter Jeun Ho Tsang:

I feel you love it, pritesh. She now has the ability to say that she's a man working in fashion and, at the same time, she's doing your laundry whilst you build your fashion tech empire. You know, I think that's the way to go. So my final question is where do you believe the future of retail tech is heading?

Pritesh Kanani:

I very, very strongly believe that the fashion industry is going to be deeply automated and digital In five to ten years. We're going to have centralized platforms for ordering, displays, monetization and end-to-end manufacturing. A shopper will come in, get themselves sized just in a matter of minutes, they will have a profile on their Google account and they could get something delivered to their door within a matter of hours which would be custom manufactured or reprinted, even to say to make such bold claims. And on the backend, everything would be digitized. There would be data patterns, there would be all the AI would go in and models. I do see such similar groundbreaking advancements at least one of them coming in the next five years. That would tremendously change, like once as a one-stop shop for the fashion industry.

Peter Jeun Ho Tsang:

Let's see if it comes true or not. Then, pritesh, I want to finish off this conversation with a quick fire round of questions. I want to finish off this conversation with a quick fire round of questions.

Pritesh Kanani:

So the first answer that comes to your head. Are you ready? Yes, as ready as I can be, and always ready. Let's get started then Fashion.

Peter Jeun Ho Tsang:

Should it fit like a glove or fit to comfort? Fit to comfort AR VR EXO or IRL EXO. Engineering or fashioneering? Fashioneering One vital skill to make it as a computer scientist in fashion. Just code learn programming. Learn code. Learn programming Easier said than done, best tip for wannabe fashion tech entrepreneurs.

Pritesh Kanani:

Just learn how to sell. Go out there, go out on the field and just design things yourself like solve end-to-end problems yourself.

Peter Jeun Ho Tsang:

Thank you so much for this conversation, Pritesh. Thank you, Peter, for having me.

People on this episode

Head Shot

Peter Jeun Ho Tsang

Host