Amazon’s unveiling of its cashierless grocery store, Amazon Go, spurred a global arms race for AI-powered retail experiences. In fact, retail AI startups raised over $1.8B between 2013 and 2018 and the latest investment in the space, in Trax Image Recognition, came to $125M in Series E led by Boyu Capital. NewtonX conducted a global survey with 500 leaders in the AI for Retail space across ten countries in order to gain qualitative and quantitative insights into applications, growth, and investment in the space. This survey included executives at large (10,000+ employees) companies that have not invested in AI-powered retail experiences as well as pioneers at AI for retail startups.
In-Store Visual Monitoring: How AI Plans to Kill Shoplifting
Shoplifting and other retail fraud cost U.S. retailers almost $50B in 2017. The most common attribution of this theft was to customers, but 30% of shoplifting was also committed by employees, and another 20% was due to administrative errors. Reducing shoplifting was a priority for 54% of executives interviewed by NewtonX, and was cited as a primary motivation for developing AI retail tools by 66% of executives at retail AI startups.
The AI tool that could most significantly reduce shoplifting — and also reduce reliance on in-store employees, in particular cashiers, — is in-store visual monitoring technology. This is what Amazon Go uses: computer vision technology and sensors on shelves keep track of what customers pull from the shelves and charges their credit card when they exit the store. Employees at Amazon Go spend the majority of their time restocking shelves, not monitoring security or doing customer checkout.
We’re now two years out from the launch of Amazon Go, and it’s no longer the only cashierless/ AI-powered grocery store. Shufersal Ltd., Israel’s largest grocery chain already partnered with a startup called Trigo Vision Ltd. that uses computer vision on ceiling cameras to identify items in customers’ carts. The chain plans to eliminate cashiers in its 272 stores.
In China, cashierless stores are significantly more widespread than in the U.S. BingoBox, a startup with $80M in funding, for instance, has opened up 200 cashier-less convenience stores across 29 cities in China. While the technology in the stores is less sophisticated than the tech in Trigo stores or in Amazon Go, the company is actively moving toward using computer vision to allow users to make automatic purchases. Additionally, the “Amazon of China,” JD, announced a plan to roll out hundreds of cashierless stores that will combine facial recognition, Radio-frequency identification (RFID), and QR codes, and has already launched five such stores.
These stores not only reduce the ease with which customers and employees can shoplift, they also reduce the likelihood of administrative errors since all inventory and stocking is data-driven and captured by the store cameras. Naturally, eliminating shoplifting is by no means the sole motivator for cashierless stores, but it is a significant one. As is, cashierless stores still require human labor for restocking and customer service, so the cost-savings come primarily from a 20% staff reduction and from billions of dollars saved in inventory recuperation.
A New Race for Retail-Focused Computer Vision Systems
Amazon Go now has four locations in two major American cities, and the e-commerce behemoth plans to open 3,000 more locations by 2021. Convenience stores are likely to be hit the hardest and fastest by this roll out — and some startups have emerged to replicate the cashierless, AI-powered systems behind Amazon Go for small and medium sized retailers, who are also, incidentally, the ones hit hardest by shoplifting.
One such startup called AiFi develops a suite of AI retail tools very similar to those used by Amazon Go (computer vision and shelf sensors), but is designed to track up to 500 people at once — a much larger number than Amazon Go is equipped for. The company has been piloting the technology with numerous retailers.
Another California-based startup called Standard Cognition has also developed autonomous checkout systems, and has partnered with a Japanese consumer goods supplier to make 3,000 stores cashierless ahead of the 2020 Olympics.
While autonomous, AI-powered retailers will likely become mainstream in China and Japan before they do in the U.S., Amazon Go’s expansion will push other American retailers to compete. Many of these retailers will rely on third party technology from Silicon Valley startups and other tech companies such as Microsoft and IBM.
In the end, the face of retail will shift to be increasingly automated and streamlined, with cost-savings arising from a decreased workforce and a significant reduction in shoplifting. As AI-powered retail technology becomes more affordable, small and medium sized businesses will also adopt autonomous checkout.