BackFashion AI gets a booster shot from pandemic
Indian startups TryNDBuy and AskSid are helping global fashion brands pivot as buyers shift from bricks to clicksAskSid’s AI engine derives consumer insights for clients based on data collected from chatbot, mail and other such interactions
PremiumThe top 20% of fashion brands are taking the lion’s share of profits globally, while the rest are struggling to survive, according to McKinsey analysts, who released their annual State of Fashion report last week. Overall, it estimates that profit will shrink by 90% and revenues by 27-30% this year, as fashion comes under discretionary rather than essential expenditure. But the worst hit are brands and retailers overly dependent on offline stores as the shift of consumers to clicks from bricks persists even after lockdowns are lifted.
Smart use of digital sales channels and AI tools for everything from customer engagement and personalization to market analytics and virtual tryouts were already starting to separate winners from laggards before covid. Now, they have become crucial for survival as McKinsey projects bankruptcy for over 40% of brands, and hard times for another 40%, in the year ahead.
In such an environment, AI startups with a fashion and retail focus have come into the spotlight. If they had to patiently persuade brands to run pilots and make deals earlier, now they are much sought after.
Virtual tryouts
Gurugram-based TryNDBuy, which was in talks with showrooms and online marketplaces to deploy its computer vision-based virtual tryouts before the pandemic, has now gone live with brands such as Jack & Jones and Vero Moda, besides being on track for deployment on Zalora, a leading fashion e-tailer in Southeast Asia.
Bengaluru-based AskSid, providing conversational AI and analytics to fashion, retail and the consumer goods industry, had already gained traction before the pandemic, but has seen a big spike in revenue this year along with global expansion. “For the period April-October 2020, compared to the same period in 2019, we have seen a 50% jump in billed revenue," says Sanjoy Roy, co-founder and CEO of AskSid.
Roy, who headed digital business for IT services firm Mindtree in the Asia-Pacific region, started AskSid in 2017. But it was years earlier, when he was doing an MBA in the Netherlands, that he first encountered the problem that the startup addressed at the outset. His wife Dolly wanted to buy tights to keep herself warm. It’s not something a traditional Indian is used to wearing and the couple was flummoxed by a selection of 200 tights from a recommended brand. Finally, they went to a Wolford store where a sales associate guided Dolly in choosing tights appropriate for her.
AskSid’s AI bot initially tried to replicate the salesperson’s process in guiding a buyer. And it was Wolford that became one of the first users of the automated tool to help with its expansion.
There were thousands of chatbots flooding the market by 2017, but few had the capability to drill down to a meaningful level of digital interaction with customers. AskSid’s fashion and retail focus, with better understanding of the domain’s semantics, made it more engaging.
The year of the pandemic has expanded use cases manifold. AskSid is now in 23 countries, supporting 15 international languages, and also expanded into FMCG brands, including a top paints MNC. More than that, it no longer describes itself as a “chatbot company", because that’s now just a front-end interface with the customer. The real action is behind the scenes, where its AI engine derives consumer insights for clients based on huge amounts of data being collected and parsed in chatbot, mail and other digitized interactions.
For example, it recently advised a client to go ahead and enter the Norwegian and Australian markets because of the kinds of enquiries from potential customers tracked online. For a brand selling a whey protein product, it prompted a switch in marketing strategy because customer interactions indicated interest from people with ailments such as arthritis and diabetes, not just bodybuilders who constituted the main target earlier.
“Because our chatbots are able to run these deep and diverse conversations, consumers are opening up and expressing their needs in a descriptive format. That allows us to derive insights from natural language interaction points and give those to the brand," says Roy. “Clients tell us they had a good idea about customers who wanted to buy from them; what they didn’t know much about were customers who decided not to buy from them. That’s the gap where our insights are bringing additional elements."
Critical tools
Data and analytics have also become critical for agility in tracking shifts in demand across geographies, categories and channels as a second wave of covid sweeps the world. While AskSid dives deeper into retail analytics with AI, TryNDBuy is dovetailing into a major pandemic-induced shift to buying things online, including fashion products. Earlier, these were limited mostly to standard products and trusted brands, but computer vision tools for virtual tryouts are expanding the repertoire of what consumers feel confident about ordering without going to an offline store, especially when returns are easy.
For example, when you go to the new arrivals section on the Vero Moda site, you get a ‘Try Now’ icon at the top right of any of the catalogue images. Clicking on the icon opens a virtual trial room where you get an option to edit the body shape in full-screen mode. You can set height, weight, bust, waist and hip sizes, as well as body shape, skin tone and hair style. You could even see how the same dress would look on different body types.
TryNDBuy, which has backing from international VC SOSV and Taiwan-based MOX, has received nine US patents around its computer vision modelling methods, says founder and CEO Nitin Vats, who was earlier a cryptography researcher at Microsoft. MOX is also helping the startup expand into Southeast Asia after deploying its product in India, where a leading fashion portal is expected to go live with it after plug-ins have validated the product on brand websites.
TryNDBuy is also working on a virtual makeup solution, because trying on lipstick, for example, in a showroom has become fraught with concern over catching covid.
Two Chinese apps, Youcam and Meitu, provide virtual makeup on real-time video, but recent curbs make them unavailable in India. L’Oreal also has a ModiFace solution but it’s limited to the brand’s products, and “not so good", according to Vats.
“Our software can convert any catalogue image of clothing into how it looks on 200,000 body shapes overnight in an automated way," says Vats. The main challenge the startup faces is getting the attention of potential clients who have been bombarded with such a plethora of virtual tryout solutions that TryNDBuy sounds like just another one of those limited, plasticky types. A tie-up with Technopak Advisors helped it demo the product for brands and now deployment by some of the leading ones has built momentum.
Malavika Velayanikal is a Consulting Editor with Mint.
Unlock a world of Benefits! From insightful newsletters to real-time stock tracking, breaking news and a personalized newsfeed – it's all here, just a click away! Login Now!
