There was initial scepticism about AI in retail, but it is now generating considerable buzz, with companies realising that the technology is here to help, as Scott Thompson reports.
The AI in retail market is set to hit $27.2 billion by 2025, up from $712.6 million in 2016, according to a recent report by Research and Markets. This will be powered by a booming mobile sector and the emergence of AI-based startups catering to omnichannel retailers.
It’s all about making life easier by automating tasks, sorting through data with superior levels of speed and accuracy, and making connections between data to enhance the online and in-store experiences.
“Fields such as machine learning, natural language processing, and robotics have immense potential to help all types of retailers – indeed all types of businesses,” says Prasit Ghosh, director of business transformation specialists, AlixPartners. “We believe that with the right understanding of AI’s business implications, retailers could use these capabilities to add tangible value to their businesses.”
Who is doing what?
This could be achieved in three key ways:
- Differentiating the offer: A good example is Tommy Hilfiger, which uses AI tools to track real-time fashion industry trends, ongoing customer sentiment, and emerging themes in patterns, silhouettes, colours, and styles.
- Personalising the customer experience: Nike has recently invested in Reflektion, a startup that lets e-commerce sites adapt content in real time to each shopper.
- Achieving cost leadership: Amazon’s acquisition of Kiva Robotics Systems to automate its warehousing operations has paved the way for traditional retailers to deploy similar capabilities.
In the UK, the ShopDirect Group was an early adopter of AI. Another leading light is Germany’s MediaMarktSaturn Retail Group, which is working with a number of startups via its accelerator programme, Spacelab.
“AI makes technologies smarter and more human, because they learn to interact with us,” says Martin Wild, MediaMarktSaturn’s Chief Innovation Officer. “This is very important for retailers to make customer journeys as intuitive and simple as possible.
“That is why we are also working on AI for our assistance robot Paul, which already accompanies customers to desired products in four Saturn stores and makes small talk at the same time.”
China’s Alibaba uses deep learning to offer clothing design recommendations, while AliPay taps facial recognition for payments. Within China itself, a range of facial recognition systems, including those provided by Face++, are already in widespread use, allowing customers to ‘pay with a smile’ – the unique movements of a person’s smile being used by the technology as payment authenticators.
In addition to robotics, Amazon is trialling Amazon Go, its cashierless concept store, which uses computer vision and deep learning to detect automatically when products are taken from or returned to shelves, and keeps track of them in a virtual cart.
Other notable players include H&M, which has launched a chatbot on messenger app Kik, allowing customers to see, share, and purchase products from the catalogue.
Clothing outlets Uniqlo, ASOS, and France’s La Redoute have each partnered with Visenze to enable shoppers to take a photo of an item in store and search for it online. Meanwhile, Ahold USA is testing a robot made by Simbe Robotics that patrols store aisles to check shelf stock and tags, and scans for floor hazards.
Suits you sir, madam
Outdoor clothing specialist North Face has deployed IBM’s Watson software to help consumers determine which jacket is best for them, using variables such as location and gender. For example, hiking in Switzerland in November and commuting in New York in March will deliver different results.
Meanwhile, Mac Cosmetics has rolled out an in-store augmented reality mirror that uses facial tracking technology and 3D rendering to enable shoppers to try different makeup options virtually, mimicking an in-store makeup artist and saving valuable time.
Internet of Business recently produced an in-depth report on similar moves by Vodafone, Mango, and others, to create digital mirrors and smart changing rooms, to upsell other garments and accessories and make the most of shoppers’ time in bricks-and-mortar stores.
These technologies can also be linked with stock control and supply chain systems, and with smart watches worn by in-store staff to create an integrated retail system fuelled by AI and customer data.
Another recent Internet of Business report explored how shoppers are demanding new AI, VR, and AR experiences in their quest for closer relationships with their chosen brands.
Meanwhile, our in-depth report on PAL value chains explored how companies such as Adidas are using AI, on-demand manufacturing, and robotics to create small ‘speedfactories’ that can make a pair of trainers to order locally, bypassing monolithic global outsourcing systems completely.
Laggards and legacy issues
Yet for all the trailblazers online and on the high street, many others have yet to see the opportunity. One of the perennial challenges they face is prioritisation, observes Martin Newman, founder and chairman of multichannel e-commerce consultancy, Practicology.
With the majority still requiring a business case for all new developments, AI might not be seen as a priority. In some cases, the opportunities it presents have also not been widely implemented, meaning that there is a lack of available data to prove the upside and the ROI.
“In that scenario, other developments will be prioritised ahead of AI and machine learning,” says Newman. “In addition, many retailers still have to contend with legacy technology. Therefore AI seems like a ‘nice to have’ – when in reality, it may present an opportunity to expedite developments. It can certainly drive efficiencies, and it is a sales conversion driver.”
Adoption is generally at a very early stage, admits Gorazd Kert, director at AlixPartners. “Most retailers are in the discovery phase or are considering the option, but not actively using these technologies.
“There are several reasons for this, including legacy infrastructures. AI-supported activities require ingestion of vast amounts of data, so infrastructures need to be agile and scalable. Sophisticated AI systems also require smart structures like AI-defined infrastructure (ADIs) and cloud-based networks that can quickly expand, based on business needs.”
Budget is another issue. “While AI can help reduce costs in the long run, it typically requires significant financial investment at the start, alongside time and resources,” adds Kert.
Upward trajectory
Nonetheless, there has been steep growth in AI investments and acquisitions since 2016, says AlixPartners’ Ghosh, adding that he believes this trend will continue.
Practicology’s Newman reckons that more retailers will begin to use chatbots to drive conversion, while also creating economies of scale and reducing customer service costs.
AI bots can manage many times the number of chats with customers than a human being is able to, and the quality of feedback can be more relevant, he argues.
He adds that conversational commerce is now getting serious traction with tens of millions of consumers having purchased or been given Amazon’s Echo (Alexa) or similar devices from Google (Google Home/Assistant) and Alibaba (TMall/Genie).
“This is heralding a move towards voice and a gradual move away from touch screens, which lends itself even more so to AI and machine learning,” he says.
“AI is already being used in payment and risk management. For example, as AI solutions gather intelligence on customers, they can enable some customers to bypass passwords and other steps in order to improve their experience.”
Ghosh concludes: “Some of the recent successful AI trends will continue to find traction in the next couple of years, but more advancements are on their way that will make this space even more exciting.”
Internet of Business says
New technologies, combined with economic stresses and changes in customer behaviour, have combined to make life tough on the high street and main street for traditional retailers.
The poorly differentiated mid-market is being pulled apart by niche, boutique, personalised retail on the one hand, and by mass-market, low-cost, short-production-line offerings on the other. At the same time, online shopping has transformed the whole retail experience, forcing bricks-and-mortar brands to get creative with multi- and omnichannel shopping experiences.
In this battle for shoppers and customer loyalty, AI can certainly help, combined with other technologies – as the above examples demonstrate. But the issue for many retailers is simply expressed: cost. In this sense, retailers need guidance and to know all of the available options in their quest to remove unnecessary friction and keep the customer satisfied.
Coming soon: Our Internet of Supply Chain conference.