For suiting company that sells online directly to consumers, hiring people to measure their clients for each suit seems a painstakingly inefficient process. For one, it is prone to inaccuracies—not everyone will have the experience required of a master tailor. And human error, however slight, can affect even the most experienced tailors. Secondly, the concept is not scalable, limiting the growth model for the business. Moreover, it is a waste of resources, time, and money, all of which could have been reinvested back into growing the company in more effective ways.
Using the latest in machine learning technologies and AI, several enterprising custom suiting companies are looking to solve sizing in just a few clicks. But does it work? Hangrr USA co-founder Rishabh Khandelwal seems to think so.
Rish, as he prefers to be known, is no stranger to the world of start-ups. Born in Mumbai, India, he was a Bollywood actor and voice-over artist from the age of eight (he played the part of the childhood of Gandhi for television) before creating micro-volunteering platform Troopp in 2011 as part of a pro-bono project. Though no longer in operation Troopp, which connects relevant skills and expertise to 3.3 million non-profits across India, ended up being one of the largest online volunteering platforms globally at the time. Rish credits his experience building it for helping him create Hangrr.
The concept for Hangrr grew in both India and England, before launching with a support team in Singapore. Rish started with a vision to automate the fashion industry, and while developing his proprietary technology (an AI-powered stylist called Lenna that predicts your body measurements in 10 seconds given a few simple inputs) to change the way men shop for custom clothing online, he also decided to learn the techniques of pattern-cutting and fitting and has become proficient in both. So #NBD but he has to date custom-fitted close to 4,000 customers in retail. We touch base with him to find out a little more about AI and how Hangrr USA is making it easy for men to shop online for custom-made apparel.
ESQUIRE: What was the most significant challenge you faced when developing a custom suiting online business powered by AI?
RISHABH: Surprisingly customer adoption was way easier than we thought. People were much happier with the minimum hassle involved in the order process, and since we were confidently able to back every product using our technology, they were thrilled that they could make use of this. But getting to this point has taken us years of preparation. The most challenging part was definitely building up our manufacturing capability that could utilise the power of our technology. We couldn't just tap into existing capacities as we found them to be inefficient and rigid. The basic methodologies have only been tweaked and not radically rethought over the last 100 years. That was very surprising.
Building this ground-up involved a huge learning curve, and it took us over three years to perfect every moving element in it. It's part of the Lean manufacturing methodology fused with the insights and power of AI. We have been able to improve delivery cycles radically while maintaining details and efficiency.
ESQ: Could you tell me a bit more about the technology used to predict your customers' sizes?
RISHABH: I have worked on numerous Image processing projects in various industries since I was at IIT back in 2008. Being able to decode the uniqueness in every individual for meaningful insight always inspired me. I believed that the next generation of solutions would be personal. That would involve deciphering the humanness behind the numbers.
ESQ: Do you think AI will do away with the need for in-person fittings for custom clothing brands in the future?
RISHABH: We have seen drastic improvements from comparing the three 'professional' measurements, 'personal' measurements and AI-predicted measurements. It's common that when the same professional measures a person twice, there may be differences in the two measurement sets received. You can imagine the errors when this is executed in person because of how the measurements are based on human perception and experience.
We started by challenging this notion that this is something that can be addressed quickly via machine learning. We took this further to rethink the art of bespoke suit-making, with proven results. Part of the idea of the in-person fitting is based on the subsequent understanding of the individual's body, gait and posture. The 3D drape is set on the human form and is highly inefficient in many parts considering the advancement of technology today. We are not trying to beat a true artist here—but we are confident that this could solve 99 percent of the need for fitting issues and expectation-matching.
ESQ: Custom clothiers can often experience high return rates of 40 percent or more due to fit-related errors. What are the return rates for Hangrr USA as a comparison for how AI has impacted custom suiting eCommerce?
RISHABH: The reason why we wanted to develop something like Lenna was to conquer this problem. We had earlier dabbled with some A.R. solutions, we did free home-trials for some time and finally, we were able to solve this problem to a large extent via artificial intelligence. We had been trying to develop this technology for over five years now, long before it was the 'in' thing to do.
I don't particularly think that return rate alone encompasses the problem here. We like to call this an error rate in our algorithm where we add an additional layer of feedback to fine-tune it. (An error is a situation even where a customer might need some minor alteration, not necessarily return the product). This is, in essence, larger in scope than the traditional return rate you have mentioned. We have managed to keep this error rate in single digits percentage points so far. We hope as time progresses, this number can be further optimised to create immense value for people across the globe. This low error rate is also a significant factor in how we can get our costs so low.
ESQ: What are your plans to grow Hangrr USA and to maintain its competitive edge in the industry?
RISHABH: I feel there is phenomenal value in further harnessing AI across various aspects of Fashion and Lifestyle, and we focus heavily on the technology that we are building. We hope that we create a mass-customised world from today's mass-produced era. Not only is it more personable for an individual's style and personality, but it is also infinitely more sustainable for the environment. It is how life should be—the way you like it when you want it.
Visit the Hangrr website for more information.