The impact of Artificial Intelligence on retail.
Meticulous Research® predicts that artificial intelligence (AI) in retail will grow at a CAGR of 34.4 per cent, reaching $19.9 billion by 2027. For most small to medium retailers AI is viewed as unobtainable technology, only accessible to the big end of town. Many believe it will cost squillions of dollars, require teams – preferably ex Googlers or Amazonians – and will take years to build.
However, what we have seen over the past year is an acceleration of SaaS-type platforms that provide plug-and-play capabilities without the significant investment. Giving small to medium retailers access to Amazon-like technology will, over time, level the playing field.
Put simply, AI is the ability for machines to take huge amounts of data, analyse it, organise it and then make decisions or solve problems in a similar way to the human brain.
The opportunity for retailers is to use Artificial Intelligence to improve systems and processes, reduce costs and increase revenue. Ultimately the biggest return will be where AI can be applied to improve the customer experience.
Here are the areas Artificial Intelligence is impacting, and some practical use cases.
AI has been applied across multiple areas in the payments ecosystem, from fraud to payment processing. In an environment where COVID-19 has forced us to rethink the implications of queueing and provide customers with a greater degree of speed and convenience, contactless payments will see huge growth. One of the most exciting case studies is Amazon Go in the US. Amazon Go is a cashier-less, cashless store. Shoppers sign up to the mobile app, enter the store and once they have chosen their product/s they simply leave the store. The payment is then automatically debited from their account upon exit. Known as “just walk out tech”, the technology uses a complex combination of algorithms. Essentially it detects activity such as a person picking up a product from a shelf, it distinguishes between the various products, and then uses image recognition to identify an individual.
“One of the biggest opportunities for AI in retail is in personalisation and guided selling.”
Some may argue that this could result in job losses, however there is an opportunity to use this technology to free up instore staff. This would allow them to provide guidance and support to customers rather than being stuck behind a till. Amazon Go now offers the ability to leverage this technology in their own stores. It will not be long before there are a growing number of competitors in this space.
One of the biggest opportunities for AI in retail is in personalisation and guided selling. In today’s environment retailers need to have a deeper understanding of their customers to develop a relationship, increase loyalty and drive repeat purchases. Machine learning not only has the ability to understand customers’ intent, sentiment and purchase behaviour, but it is also able to predict when and what they will buy. It’s no secret that up to 30 per cent of Amazon sales are generated from personalised recommendations.
Dynamic Yield is another example of mass personalisation and its opportunities. Recently acquired by McDonald’s, it presents your menu based on previous purchase behaviour, time of day, weather and local preferences.
Dynamic Yield also powers the e-commerce solution of Lamoda (Russia’s leading online fashion retailer), which reported a $15 million uplift in gross profit in the first year of implementation and an 8 per cent increase in revenue per session. To achieve this, the company created over 160 unique visitor segments and automatically targeted each with personalised offers and messages based on purchase behaviour and product preferences.
For smaller retailers there are lots of plug-and-play options that allow you to embark on the personalisation journey. Shopify has over 150 apps that provide the ability to give personalised recommendations and upsell or cross-sell. Klaviyo, an AI-powered email platform, provides the ability to predict purchases, future
One of the most frustrating experiences for customers is the search process. Search for “yoga pants” on one of the biggest retailer sites in Australia and you are presented with jeans. This leads to a subpar user experience and increases the chance of customers leaving your site. Over time our searches are becoming more specific and, as technology improves, we are shifting from one-word searches to typing the way we think (natural language). For instance, “red party dress” would be “a red party dress for a 40th birthday party”.
This is where artificial intelligence makes perfect sense. With potentially thousands of items in retailers’ catalogues, or even 100, it is impossible to manually tag every product with the right information. Manual tagging requires a huge effort and is extremely inefficient. With visual recognition and tools such as Okkular AI, which is able to recognise style, patterns, shapes, lengths and apply automated tags to products, retailers can return more accurate results and increase conversion rates.
COVID-19 has also driven a huge shift in consumer behaviour. According to Australia Post, 5.2 million Australians shopped online in April – an increase of 31 per cent compared to 2019. This resulted in a huge Increase in packages being delivered.
Retailers have come to realise that logistics is now a fundamental part of any e-commerce business. Prior to COVID-19 it was seen as an add-on, but now it is core to the customer experience.
Choosing logistics companies using AI for optimal routing will drive costs down and allow retailers to offer more competitive shipping rates. Shipping costs are the number one reason for cart abandonment – in fact, Dynamic Yield estimates that $18 billion in revenue is lost per year as a result of customers abandoning cart.
Historically, advertising has been driven by creative rather than being data driven. When customers receive personalised ads, product revenue grows by 38 per cent.
Westfield shopping complex in Shepherd’s Bush, London, has recognised the value of personalised ads using AI facial recognition. Cameras can define a shopper by age, sex and even their mood, and then display digital advertisement boards based on them.
When chatbots first launched they were the most frustrating experience. Asking a few questions would normally result in a “computer says no” response. The initial chatbots were not artificially intelligent; instead they were a set of rules based on user input. If the user input was not the same as the preprogrammed text, the question or enquiry was rejected.
Chatbots have evolved since they first came to market around 10 years ago, and with the market set to grow 30 per cent by 2025 it is a clear opportunity for retailers to connect and engage with customers.
AI chatbots learn and understand things like variations of the same question, context, and emotion. The most powerful opportunity for chatbots is to know when to hand over to a real human – to understand when a customer is getting frustrated and/or is going to drop out of the sales funnel, and make a seamless transition to a real person.
AI should not be seen as technology that is going to allow the big end of town to take over the world and leave us all jobless. It should be viewed as an enabler. Investment and implementation should be committed to only if there is a clear additional benefit to the customer experience, which in turn saves costs or increases revenue.
Kelly Slessor, founder and CEO, Shop You
This story first appeared in issue 31 of the Inside Small Business quarterly magazine