Artificial Intelligence and Retail
19 February 2018
No other sector is more awash with data than retail. The real challenge for retailers is knowing when and how to apply technology to this data, to understand and deepen the relationship they have with customers. Which is where Artificial Intelligence comes in.
That artificial intelligence has moved out of the realm of science fiction and into our everyday lives is proven by the fact that millions of people ask Alexa, Siri or Google Home to answer questions and perform tasks for them every day. And that the Internet of Things and the Fourth Industrial Revolution have accelerated this uptake is also pretty much unquestionable. Artificial intelligence’s capacity to crunch colossal quantities of big data in real-time in order to mimic human interactions, and above all to learn and adapt to deliver smart personalisation, are the factors that are, quite rightly, exciting technologists.
The Director of Business Development at Intel’s Artificial Intelligence Products Group, Azadeh Yazdan, puts it clearly:
“We live in an era where a tremendous amount of data is being generated every day and the numbers are only growing. Just collecting or having access to larger data sets, however, doesn’t lead to improved business results – being able to extract meaning from big data to solve problems and be productive is the key to success.”
A recent blog from retail experts GDR articulates it well: “This sort of insight-based problem solving [is something] that data-driven retailers are particularly good at, and as more formerly pure-play e-commerce retailers open physical stores, they are using their inherent grasp of technology and lack of legacy overhead to outmanoeuvre incumbents.”
The IBM Institute for Business Value’s recent report Thinking like a Customer uncovers many thought-provoking stats including that 91% of retail executives familiar with cognitive computing say it will play a disruptive role in their organisation; 94% intend to invest in it.
So the technology suppliers and commentators are very excited, but we’d expect no less. Let’s take more of a look at what AI might mean specifically for the retail sector
AI in action
The term Artificial Intelligence is generally applied when a machine mimics cognitive functions that humans associate with other human minds, such as learning and problem solving. What are some of the applications set to thrill the retail sector?
One area that is inarguably at the sharp edge of “pretending to be a human” is chatbots, where computer applications hold conversations with humans in real-time via text. Chatbots ‘learn’ from each conversation to provide increasingly useful and relevant responses.
Today there is big money in getting recommendations right and curating appropriate content – digital disruptors like Netflix for example see huge value in accurate recommendations for ‘locking in’ customers.
Natural Language Processing
AI’s ability to accurately process huge amounts of unstructured data is opening up all sorts of doors for retailers. A customer could for example type (or, with ever-improving speech recognition, say aloud) “I’m looking for something smart to wear for a friend’s wedding where the theme is navy. The wedding’s in October and I know the church so I want something warm, but don’t mind if it shows off my great legs. I’m in my forties, I’m a size 12” and the AI will return relevant, personalised recommendations.
Real-time insights from smarter, connected stores
IBM’s Watson is just one of the platforms driving the ability for retailers to put insight in the hands of the store staff to improve customer interaction and transform the in-store experience. Cognitive computing and rich data combine to help retailers streamline operations and infuse intelligence into merchandising and supply networks.
Facial recognition and behavioural profiling
Recognising customers (or, for that matter, known shoplifters) has a lot of value for retailers, especially when this can be combined with behavioural recognition and store journey insight. Keeping up with colossal amounts of data to derive insight from it (in real-time, if the data is to be used to serve up personalised offers) needs cutting-edge machine learning.
Fraud identification and prevention
AI will help spot patterns, filter ‘non-human’ users, and identify new trends or ‘scams’, with far greater speed and accuracy. As the move away from the use of name and password for identification accelerates, AI can help provide innovative but completely secure methods of individual identification.
An insightful panel debate at NRF ‘Robotics and AI: Trailblazing Technology for Future Retail’ started from the blunt fact that retailers are under immense pressure to cut costs, which often means instore staff. Physical retail still a great experience but taking staff out of the store negatively effects this. If you can get robots and AI to cover back-room duties, you can free up your staff for more customer interaction. AI is key to manageable robotics, particularly when they might interact with shoppers or be on the floor at the same time – for example to check inventory.
Augmented Reality and Mixed Reality
The consumer appetite for augmented reality applications is a hungry one, and brands are keen to tap into this. Breaking down the barriers between virtual and physical reality can make for a compelling, differentiated engagement, but needs artificial intelligence and big data to ensure a meaningfully personalised experience. Microsoft’s mixed reality technology HoloLens is being used by Volvo Cars, for example, to reimagine how people experience car buying.
IBM’s Watson is impacting on stock management, Watson can monitor a disparate range of variables from weather and purchase rates through to wider consumer trends, enabling smarter supply chain management to right size inventory and minimise stocks-outs. (IBM’s commitment to AI is nothing less than existential – they acquired The Weather Channel as part of their mission to use AI knowledge to support supply chain and predictive buying.
Bringing sales and marketing departments together
AI is much feted by retailers as they sees it potential to automate and ultimately revolutionise larges tranches of work, and improve both conversions and forecasting. AI applications can analyse inbound emails and suggest appropriate next actions based on past behaviours and conversions. More widely than just inbound, using data collected to predict ‘intention-to-buy’ signals on websites and marketing collateral enables sales teams to focus resources on the hottest leads – ie the ones most likely to convert.
The Salesforce ‘Big Ideas Session’ at NRF this year had a great title – ‘You Can’t Spell Retail Without AI’. They see AI as nothing less than central to the future of retail and as the arena which will, as their blog pointed out, “separate the leaders from the laggards”.
The ethics of Artificial Intelligence
NRF 2018 is inevitably still very much on my mind as I write this, and I’m reminded of a very thought-provoking panel debate with some of the very top leading thinkers in the AI field, that discussed the emerging field of AI ethics. If you’re of a philosophical bent and have a spare 45 minutes you’ll probably enjoying watching the video. Some of the considerations discussed include the following:
Transparency is possibly the most important consideration in the ethics of AI – for example, being clear to your retail customers that they are talking to a chatbot rather than a human.
The authenticity agenda in retail is gaining ground – is this in conflict with the AI push?
Can we use AI to avoid either accidental or purposeful discrimination? Are we sure that the data we are using is clean and unbiased, and then is it easy for us to explain why the AI algorithm acted in a certain way?
Ethics-by-Design – problems can be avoided by framing questions early that help the design of the interface and the user experience. ‘Ethical’ conversations have to happen at the very beginning, and aren’t something that needs to be left to the designers and developers. In other words, AI can’t be left solely to the AI people.
Insight vs self-insight – retailers and manufacturers will now have more insight on consumption than individual consumers do. This allows both monitoring and possible regulating psychological behaviour with regard to preferences and choices. We can know for example a lot about the behavioural disposition of an individual consumer. At what point should a company decide that it can still market alcohol to an alcoholic? Credit card companies took a long time to understand that they shouldn’t market to certain people. The challenge is still: who makes that decision?
The ‘known unknowns’– we must bear in mind that systems are interdependent, so we can’t necessarily predict how AI’s spread may effect society – we probably wouldn’t have thought 15 years ago that social media might threaten the exercise of democracy, for example.
While some of these topics may feel very high-level, the point is they need thinking about now, to get right, rather than having to ‘mend’ them later…
Commitment at the highest levels
Here in the UK, the Government is committed to building on the reputation the country has (through such AI luminaries as Alan Turing) as pioneers and innovators in the AI field. But this isn’t about national pride or nostalgia: it has been estimated that AI could add an additional £630bn to the UK economy by 2035.
In its report ‘Growing the Artificial Intelligence Industry in the UK’ the Government articulates its vision is for the UK to become “the best place in the world for businesses developing and deploying AI to start, grow and thrive, and to realise all the benefits the technology offers.”
The Chairman of the Parliamentary Sub-committee for AI, Lord Anthony St John, will be speaking at our Retail Tomorrow event in late February. We look forward to his insight around the specific opportunity for retail.
The intelligent retail future
As 2018 starts AI has been everywhere, including sessions and discussions at NRF and at Retail Tomorrow. But it isn’t a ‘trend’ for 2018; it’s here to stay. The Internet of Things, the Connected Store, the Fourth Industrial Revolution … these are all driven by, and are reinforcing the importance of, AI in a data-centric future. Retailers neglect it at their peril.
I’ll leave you with a parting thought from the IBM Watson team:
“The retail future isn’t about products – it’s about possibilities…”