M&S, Co-Op and Dunnhumby discuss the opportunities and risks of AI in retail
Three of the largest companies in the UK to be working on AI in the retail sector, M&S, Co-Op and Dunnhumby breakdown the reality of what the technology means for shoppers.
It’s not often that you get the chance to hear from three major companies operating in the retail sector, discussing what their plans, opportunities and challenges are for the use of artificial intelligence (AI) and machine learning (ML) technologies, given the competitive advantage their use can bring and the risks involved with getting it right.
However, this week at Microsoft’s Future Decoded event in London, representatives from M&S, the Co-Op and Dunnhumby sat on a panel and discussed openly how their respective businesses are jumping on the AI hype and considering its use cases.
M&S and the Co-Op are two of the largest retailers in the UK, whilst Dunnhumby started as a data science firm and was behind the first loyalty card for Tesco (another large UK supermarket chain). The trio of firms discussed openly not only their plans for how they’re investing in AI, but also the challenges around employee engagement.
We are seeing a lot more demand from retailers really adopting this, not just experimenting. And they’re really thinking about the strategy around where they apply this to which use case, where they apply this to particular problems, thinking through and building the right roadmap.
Not just looking at how they can drive share of wallet, getting customers to buy more, but also driving efficiencies in processes, making things quicker and easier, a bit more automated. Also, for driving new revenue streams, data is a great asset.
Simon Locke, head of IT at the Co-Op, said that the company is starting to use AI at large to introduce efficiencies into back-end processes, and is also experimenting with how it can be used for customer engagement. He said:
We started our journey in RPA, automating a lot of our back-end processes, recruitment processes particularly. We are starting to scale that now. We are experimenting a bit with video recruitment at the moment. We are experimenting in the stores, with queue management and looking to predict where queues are building up and to get people onto the tills. There is some experimentation happening around checkouts, looking at suspicious activity.
A fair value exchange
Clements added that the use cases will be different for every business and that retailers should be thinking about what their burning platform problem is. For example, he said that Dunnhumby is currently focused on perfecting promotions for retailers, where AI and science can add “massive value”. Clements said that this can be around working out cannibalisation, the best mix of promotions, but also reducing the amount of time spent on planning a promotions strategy. He argued that data science and AI can “add massive value” in this area.
The topic of bringing value to customers dominated the discussion. Clements also noted that there are different currencies to think about - not just personalised offers, but also being helpful, saving time, and making things more convenient for customers. He said that “these are all things that customers will recognise and that’s going to be a great value exchange”.
Paul Dasan-Cutting, Innovation Product Owner at M&S, explained that the company hasn’t always explained the value exchange with customers, of handing their data over in order to benefit from a M&S loyalty card. He said that AI can help with this:
We recognise that there was a need to create a loyalty scheme. So we did that. But we didn’t ask ourselves the question, what do we want that to actually be? What is the value to the customer in collecting all that data?
We have got to a point where we are sort of offering some personalised offers. But actually we aren’t applying the right level of analytics and insight onto that to provide you with a true personalised offering. We are looking at how we can change that, to incentivise customers to sign up, so that it makes sense for them to do that. Because we can offer them an experience that adds value to them in exchange for that information. Then it’s a shift about delivering value to the customers.
M&S is also exploring how AI and ML can be used to reduce waste. Dasan-Cutting said:
For me, personally, it will be about how can we use technology to reduce wastage? We know when a product is going to go mouldy or out of shelf life, how do stop that being sold and also sell it for the full price as long as possible? How do we get it to the right customer at the right time? That’s where this technology can add value.
Meanwhile, Locke said that the Co-Op is wanting to figure out how data can be better used to boost employee retention, or at least introduce efficiencies into the recruitment process. He said:
We have a massive problems with recruitment, or retention. We have 50% turnover of staff in the year. So we need to make our recruitment, retention, on-boarding processes a lot more efficient.
If you’re looking at the front end of the store, it’s about how we marry up the digital parts of our offering into the human touch. As a convenience retailer, we are in the heart of every community and it’s around providing that human touch.
Locke was also keen to discuss the data challenges facing the Co-Op. AI and ML are clearly useless without data to feed the algorithms and systems. And as most will know, organisations have historically been terrible at managing and organising their data in a helpful way. He said:
One of the challenges we have with data is that the co-op is such a broad organisation in terms of the number of businesses we have. Whilst we have a data strategy, we haven’t yet brought all those things together. Do we need to bring every piece of data together to leverage it? It’s an interesting point.
Where we are doing that, it does make a difference. Last week we launched a new Co-Op app, it providers personalised offers to individuals, and that will help drive people into the stores. We won’t be sending you offers of things you don’t normally buy, we will be sending you offers of things you do normally buy. It’s not a huge starting point, but it’s making a difference. But how we collect the data and making that really transparent is really important.
Unsurprisingly, the firms were also quizzed on their attitude towards AI and the impact it may or may not have on peoples’ jobs. It will come as no surprise that research around automation and the use of AI has pointed to the potential for the huge displacement of jobs and the requirement to quickly reskill the workforce, so as to manage the change. However, Dunnhumby, M&S and the Co-Op all said that the key is engaging with the workforce and ensuring that they’re involved through the process of introducing the new technologies.
From Dunnhumby’s perspective, Clements urged companies to focus on the benefits to employees and to educate them on those benefits. He said:
I think this is about a big transformation in our industry. One of the most important and biggest challenges is change management. And actually working with organisation and really thinking through, what are the processes in the business? How can we improve those? And getting buy in from the people that are used to those traditional processes around new processes?
How they can improve their jobs by freeing up time, so they can spend more time on added value activities, more time on innovation, discussing other things with suppliers. Do test and learning, improving as you go. When you see something is successful and you get buy-in, you can scale it very quickly.
Dassan-Cutting agreed and argued that the focus needs to remain on improving the employee and customer experience. He added:
I think it’s about perception and branding whatever you’re putting in the hands of your customers and colleagues. How do you get your workforce, your customers, your colleagues along the journey with you and try to get them up to speed, to an understanding that this isn’t a threat? This is going to make their jobs easier. This is going to make the customer experience better.
And get that feedback loop working - where they can say that this isn’t working and push changes through. If you can get them bought into that feedback loop, it’s going to make it so much easier to adopt AI and anything other technology.
Finally, Locke argued that the use of AI and ML will actually mean that employees will spend more time on value-add work, rather than managing back-end processes. He said:
When we built our new store recruitment system, we actually had three store managers with us full time on the project for six months. And that really made a massive difference in terms of its buy in from stores.
I think the way we operate as a business too, we account for every single minute of a person that works in the store. And we want those minutes to be in front of customers, we don’t want them spent in the back-office doing processing work. I think if the story is told in the right way, it can be a huge advantage.