From driverless cars to grocery stores with no cashiers, robots in factories and highly tuned algorithms which can outpace human office workers – the robotic revolution is well underway: at least if the headlines that ‘the robots are coming for our jobs’ are to be believed.
In 2018 the World Economic Forum predicted that machines and automated software will be handling half of all workplace tasks within seven years. As well as potential job losses, however, the report also suggested that innovation in areas such as artificial intelligence (AI) and robotics could create more jobs than it threatens.
In retail, the disruption is being led by online behemoth Amazon as it increases spending on new technology, from robot-enabled warehouses to its ‘no line, no checkout’ Amazon Go convenience stores, in an effort to cut costs.
In response, its competitors are fighting back with new technologies. The world’s largest retailer, Walmart, is opening data-driven ‘intelligent retail stores’ and bringing on board thousands of robots to do lower-level jobs such as scrub floors, scan boxes, unload trucks and track shelf inventory.
Fellow US grocery chain Kroger is trying to fight the ‘Amazon effect’ by working with Microsoft to build ‘connected experience’ stores. These will give shoppers personalised deals, either via their phones as they walk inside or on screens mounted on the shelves.
Another retailer leading the way in automation is China’s JD.com and its Silicon Valley research centre. It is implementing everything from robotic arms for picking, to automated delivery drones, to AI and big data technologies on an immense scale, with just under 7,000 distribution centres providing same and next day delivery.
Fashion flagship stores including Zara and Saks now boast smart or ‘magic’ mirrors which suggest different styles or tell you what other stock or sizes are available. According to the Guardian, 59 per cent of UK fashion retailers use facial trackingto counteract theft.
Meanwhile everyday automation is being used by New Zealand retailers.
Self-checkout kiosks have been announcing there’s an “unexpected item in the bagging area” since 2006, when they were first introduced by Foodstuffs, and are now used by retailers such as The Warehouse and Kmart.
The Warehouse also offers an ‘Endless Aisles’ service, in which in-store shoppers can access the full online range at any store. In May it introduced what it calls “Australasia’s first automated click and collect tower” at its Sylvia Park store.
The five-metre tower acts as a vending machine for online orders, allowing people to retrieve their parcels in seconds. It can hold up to 300 parcels at a time which are loaded on to trays, measured for size and arranged inside the tower dependent on their height. Packages are then released to customers within seconds through smart sensor technology.
The Warehouse CEO Pejman Okhovat says the tower is part of a three-month pilot that focuses on omnichannel retail and aims to provide customers with a more seamless way to shop.Almost a third of online orders are now collected in store, he adds.
Since 2015, Auckland Airport travellers have been able to collect their pre-paid duty free from Aelia’s white crane-like robot named ACE (aka an ‘automated collection experience’). Online, brands like Cotton On allow shoppers to message chatbots.
The potential of what could soon be applied in retail is demonstrated in other industries such as banking, where ASB and ANZ are trialling physically realistic, digital AI assistants, named Josie and Jamie respectively, as well as Air New Zealand’s Sophie and Vodafone’s Kiri. These intelligent digital humans are designed to sense the emotions of the customer and respond accordingly.
Retailers who are customers of Vend’s point-of-sale and retail management platform may be familiar with Dott the AI bot. The non-gender specific Dott offers insights based on a retailer’s activity, more akin to a business partner than a voice-activated assistant like Alexa and Siri.
The retail industry is a prime candidate for automation, reliant as it is on low-skilled roles, and given that a retailer’s largest operating cost is generally its workforce.
In New Zealand as well as countries like the UK, changes to the minimum wage are viewed as biting into margins, leading many retailers to look at cutting their wage bill.
Process-driven jobs that can be easily automated are most likely to be displaced, says Dr Jessica Vredenburg, a marketing senior lecturer and assistant professor at Auckland University of Technology (AUT).
“Essentially any job that can be done more efficiently by machines than people.”
Retail employees view the growing use of technology in their jobs as both a concern and potential opportunity. In February a US study, by The Fair Workweek Initiative, found that 63 percent believe their current retail job won’t be fully replaced by technology during their lifetimes.
Many jobs do still benefit from or even require a human touch, says Vredenburg.
“For the foreseeable future at least, it will be difficult for machines to replicate soft skills such as empathy, emotional connection and communication so humans will still be needed for jobs that require this sort of relationship building.”
Vredenburg and her AUT colleague Dr Yingzi Xu are currently looking at frontline automation in the airline industry; specifically, the introduction of the check-in kiosks and how that is disrupting jobs.
Automation is “absolutely changing” the nature of workers’ roles. This has big implications, whether that’s in the airline or retail industries.
“The employee role is no longer clearly defined, and the service process no longer unfolds in a predictable manner. Employees are required to multi-task, problem-solve, and work in an, at times, physically crowded and demanding environment, such as self-service checkouts.”
Her study also looks at how customers and employees interact win this situation. “Some customers are also still hesitant to use the technology: they are unsure how [to use it], worried they’re going to make a mistake or simply don’t feel it’s their job.
“This results in potential tension for the employee who is still expected to provide customer service but in a different way, facilitating the customer to self-serve (and not just doing it for them), but being at the ready to jump in and assist should there be an issue.”
In an era of instant gratification, customers are demanding efficiency, convenience and control, she adds.
Under these circumstances, automation “can be very efficient… much more efficient than the human-based systems”.
At the same time, technology does not exist in a vacuum. It still requires humans to run it and provide support when it breaks down.
“There is no doubt that automation is the way of the future, but it is our view that the goal is to find the right balance of tech and touch. The future of customer service is one in which human and machines complement each other working together to deliver the best possible customer experience.”
Automation and AI offers the potential for “a new wave of retail experiences”, says Asa Cox, chief everything officer at Intela AI in Wellington, which specialises in AI and machine learning.
Consumers are used to receiving personalised recommendations from Netflix or online retailers, he says. “From a retail perspective, I think taking that into the real world is eminently possible.”
For example, there is huge untapped potential in loyalty card data for dynamic shopping: offering promotions, discounts and product trials to the consumer while they are in store.
“There’s a real potential for it to become a personalised shopping experience using both in-store technology and also the data that the retailers have available, because of the data we’re giving those retailers. Taking our online data and using that in the real-world retail experience.”
For retailers, the biggest “near-term advantages” are in backend processes, says Ben Reid, executive director of New Zealand’s AI Forum, which represents local and multinational tech companies involved in developing AI.
Take any large organisation, he says, and there will be many backend processes that depend on repetitive tasks which are open for automation.
“It actually frees people up to do the more interesting and the more complex stuff.”
Machine learning capability offers a way of processing all this, via data science and, increasingly, deep learning, finding patterns and processes that the human brain simply can’t see.
Retailers should look for the processes within their business that are repeatable. This delivers huge efficiencies, he says. “Those tasks get carried out very reliably and can operate 24/7”.
Historically, businesses have used structured data, such as database tables, to base decisions on. Now, however, the quantity of data being generated every day is growing exponentially, and includes unstructured data such as video, audio feeds and images.
AI is needed to make sense and gain insights from all of this, says Reid.
Yet it seems New Zealand businesses, including retailers, are hesitant about getting into the automation space.
In February PWC New Zealand’s annual CEO survey found that 84 per cent believe that AI will transform their businesses within the next five years. Yet 30 per cent have no current plans to pursue any AI initiatives.
Some of this uncertainty comes from a lack of knowledge and skills to truly understand the potential impact of these technologies, says Cox.
A lot of businesses are still undergoing cloud migration, he adds, “so perhaps there’s a bit of transformation fatigue”.
New Zealand does not have the same level of competition, or scale which in bigger markets often drive change. There’s also the perception, he adds, “that it’s all very, very expensive and hard”.
In reality ‘transformational’ technology does not have to cost millions, or take years to trial, says Cox. “You can really get a proof of concept going pretty quickly, pretty cheaply.”
With a video feed, for example, “you can get something interesting after a couple of months and $50,000 dollars. It doesn’t have to be a huge amount of money to really start understanding the value of some of the machine learning in AI to really prove to the execs, to the business that this is an area they should invest in.”
Machine learning tools like Intela AI’s Farrago allow retailers to use data in ways that previously only data scientists could: from predicting the best prices to ascertaining product demand. This data can be used in marketing campaigns and in-store promotions; it enhances the customer experience but also boosts profitability and reduces costs.
Meanwhile its machine vision tools take standard CCTV video feed and use that to collect aggregated data. New Zealand has strict CCTV laws but footage is processed in a way that removes any personal identification.
“We know where people are exactly in the store and you can actually get a much better understanding of your retail environment to create planograms, or understand how to increase the size of basket, what to do around special brand promotions and so on.”
Retailers should give AI a go, he adds.
“It’s not scary, just start exploring what’s possible, because unless people start now then they get left behind, and that’s what people should be worried about.”
This story originally appeared in NZ Retail issue 762 June/July 2019