Machine learning has a wide range of applications in retail and new technology allows it to be accessible for retailers of all sizes.
As we’ve seen, Artificial Intelligence (AI) is transforming retail. Tough conditions, intense competition and tight margins are all driving a focus on optimisation, efficiency and avoiding waste. To better understand customer behaviour and create actionable insights, marketers and retailers are harnessing big data and accessing the potential of the Internet of Things (IoT).
Vast amounts of online and offline data are being collected – everywhere from websites to in-store cameras and sensors. But the challenge is not just gathering and storing it but also getting something valuable out of it. Retailers need insights about customer preferences and behaviour, shopping trends and future predictions.
This is where machine learning and AI is now starting to play a major role. Not as android shop assistants or robot shelf-stackers, or at least not yet anyway. These technologies are about data and intelligence: enabling better decision-making, to enhance and augment the customer experience and improve profit margins.
AI is intelligence demonstrated by machines, applying a set of logical rules to make decisions in the same way a human would. Machine learning takes this a step further. It enables a machine to use available data to refine its rules and improve its decisions. Computers, with their massive processing power and ’unbiased’ view, can detect patterns that humans can’t.
Machine learning has a wide range of applications in retail. It can be used to generate highly customised and personalised product recommendations. It can optimise pricing and make dynamic adjustments. It can also predict future sales patterns and refine inventory planning.
Statista’s recent survey highlighted cost savings, increased productivity and revenue, and more informed business decision-making as the top benefits of AI for retail businesses. According to McKinsey, retail supply chain operations who have adopted data and analytics in the United States (US), have seen up to a 19% increase in operating margins over the last five years.
Major US retailers and global brands have been using AI and machine learning for some time. Amazon’s Recommendation Engine is one of the most famous examples – its algorithms are so effective that its machine learning recommendations drive 55% of sales. Netflix is another prime example, it tracks viewing habits carefully, and refines its content buying and scheduling accordingly, saving $1 billion a year through the insights generated.
Until recently, smaller retailers haven’t had the resources to implement similar technology. But now, software solutions accessible to the smaller end of the market are becoming available.
M-intelligence, developed by meldCX and AOPEN Group, is one of the first commercially-ready machine learning tools for Microsoft Azure. It’s product recognition technology enables retailers to cultivate and teach its own training models. A key advantage is that this can be done without a single line of code – critical for smaller retailers without large IT departments or data scientists on call.
M-Intelligence uses an edge processing methodology which allows transactional speeds offline while being updated and educated online. Images are captured by a human trainer, uploaded, analysed and the reference data is stored in the cloud.
This enables features such as product recognition, to reduce supermarket self-checkout fraud. For example, over time the system scans many different bags of nuts, and learns to differentiate between peanuts and cashew nuts, to a specific level of certainty. Depending on whether this is high certainty – ‘this is a cashew nut’ – or medium – ‘is this a nut?’ – the system can trigger further actions, such as ‘print label’ or ‘select type of nut’. A low certainty match might alert a human store assistant to attend.
It’s no longer only an experimental concept. The machine learning technology is now working so well that it’s going into proof of concept (POC) rollout later this year, with strategic partners such as Intel and Microsoft coming on board.
Machine learning and AI may still be in their infancy as technologies for retail, but what we’re seeing so far is both impressive and transformative. They’re already starting to revolutionise retail as we know it.
Joy Chua is currently leading the Strategic Partnerships and Market Development team at both AOPEN Group and meldCX. Skilled in strategy, marketing and communications, Joy is passionate about digital transformation especially in retail and merging the online to offline customer experience by utilising the best of web in a physical space. Learn more at www.aopen.com/au/home
