g that are available for sale, and through machine learning and visual matching algorithms, can generate realistic images of the person wearing different garments. Through the technology, users are able to customise the style, colour and pattern of the product worn by the model to their own tastes. These models have been chosen to represent different skin tones, body shapes, ethnicities, and hair types. In addition, using machine learning and visual matching algorithms, users can make changes to the style, colour, and pattern of the product, which would be reflected in the model. The company also utilised Google’s Shopping Graph – a comprehensive dataset of products, sellers, brands and inventory – to reflect how the product would drape, fold, cling, stretch and form wrinkles and shadows. In developing this technology, Google is seeking to address a number of issues related to the online shopping experience, including customer dissatisfaction with clothing fit, and significant challenges related to returns and textile waste.  Currently, US shoppers can use the technology for women’s tops from brands including Anthropologie, Everlane, H&M and Loft, but there are plans to expand this service to include more products and ranges over time. Driving in-store purchases Kelly Slessor, founder and CEO of Shop You, an AI-driven personal shopping assistant, told Inside Retail that machine learning and visual matching algorithms will eventually impact every element of retail – from customer service, to in-store visual merchandising, pricing to inventory management. She explained that her company, Shop You, tried to utilise similar virtual try-on technology about eight years ago, but it wasn’t ready-for-use at the time. She believes Google’s application is “revolutionary” – and  that it can be deployed at mass. “Virtual try-on isn’t new [and] it is also widely used in the makeup and beauty industry,” she said. “Customers need to be able to visualise what something will look like on them [and] virtual try-on is a great way to bridge the gap between online and offline shopping. It allows people to try on clothes from the comfort of their own homes, and it can help them find clothes that fit them better.” She pointed to data which showed that the virtual try-on market was valued at US$3.5 billion in 2021, and forecasted to grow to $12.97 billion in 2028. Retailers are also starting to recognise various ways to use AI – with Slessor adding that it will help to drive in-store sales. “If we can get customers who prefer in-store shopping closer to the purchase point online, driving them to the physical stores becomes easier,” she said. “Imagine the customer journey: searching for a red dress, trying it on virtually, and then adding it to my wishlist or reserving it in-store ready for collection, knowing already that it will suit me.” Robust protections Meanwhile, futurist Gihan Perera believes that Google’s virtual try-on feature will enhance the digital customer experience, benefitting both customers and retailers. He affirmed Slessor’s statement that it will narrow the gap between online and offline shopping experience, explaining that the feature is not the same as trying on clothes, but is an improvement to a “one-size-fits-all photo.” “For some customers, [virtual try-on] is good enough in exchange for the convenience, speed, product range, and other benefits of online shopping,” Perera said. For retailers, the potential benefits include saving money – as they don’t have to photograph different models wearing different types of clothes – improving customer engagement and conversion rates, and reducing return rates. Perera added that generative AI thrives on large data sets, which consumers are providing in droves. By analysing this data, the experience will inevitably improve, with customers subsequently receiving a more personalised experience. However, this new technology is not without its potential drawbacks. For instance, customer expectations might exceed capabilities, which could lead to disappointment in the experience, as well as an increase in returns. Furthermore, Perera observed that AI images have the potential to be disputed, especially when professional models and original images are used. “Customers are not uploading photographs or 3D models of themselves, so that personal data isn’t being shared,” he said. “But Google and retailers can still infer some private information, such as skin colour, ethnicity, body type, and physical measurements. There need to be robust data privacy measures to protect customer information, disclose how the data are used, and ensure secure usage.” Slessor also noted that the potential for privacy and security breaches is a major concern. With generative AI potentially paving the way for users to upload images of themselves, there’s a heightened risk of images being hacked or leaked. “An obvious fear would be virtual try-on underwear. No one wants those images leaked, so there will be concerns over data storage, but [I] believe there are ways to mitigate this, ensuring images are owned and held locally on the user’s device,” she said. “The customer journey will need to be built to ensure that privacy and security are the focus. Data protection and privacy are obviously a major issue across the entire AI landscape which needs to be addressed by both technology and legislation.” Boosting conversions The potential for virtual try-on has already been actualised in other product categories. According to Slessor, its implementation in the optical sector led to a 600 per cent increase in conversions. It has also been used in the makeup industry to great effect, leading to a reported 35 per cent increase in overall sales. “This significant uplift in sales highlights the huge potential of virtual try-on technology to boost conversions and ultimately drive revenue for fashion retailers,” she said. Beyond the fashion industry, Perera highlighted the fact that Ikea is using virtual try-on to visualise how furniture, paint colours and other items look in specific places. He also identified the fact that Mercedes-Benz has a virtual reality showroom for customers, with the tech being used to simulate how different car models, colours, and customisations look and feel. Slessor added that people who use virtual try-ons can make more informed decisions, which could help them avoid buying clothes that are a poor match, and that they would return later on. “The positive knock-on factor of this is the reduction of unwanted items that eventually become landfill,” she said. She added that virtual try-on is “the biggest advancement in inclusive technology” that she has seen – with the representation of the models allowing customers to see people who look like them.  Further, she believes that it will become more accurate as the algorithm improves.  “The images are becoming more realistic and will better reflect how the clothes will look on the user,” she said.“It is also becoming more interactive, users will be able to interact with the clothes in a more natural way. For example, they will be able to rotate the clothes, zoom in and out, and try on different accessories. “It addresses the dissatisfaction felt by individuals from traditionally underrepresented groups by providing them with a personalised and inclusive shopping experience.” An exciting moment According to Perera, the “obvious next step” is a scenario where people can upload a version of themselves, and try on clothes using a virtual version of themselves. They could also upload their wardrobe and try on clothes and accessories to match what they already own. “Of course, ethical considerations and privacy concerns must be even more carefully managed in such scenarios,” he said. He also stressed that organisations need to be transparent about the technology’s limitations, and provide customers with an alternative way to transact, if they don’t feel comfortable using the technology. Meanwhile, Slessor believes that smart dressing – the ability to buy or view items based on a variety of factors including personal data, the weather, or what’s trending – will be an interesting possibility to explore moving forward. “We are at an exciting and pivotal moment in retail, [and] education is key to embracing change and ensuring that organisations are AI ready,” she said.