As more retail businesses are integrating artificial intelligence into their operations, we take a look at the current state of play and offer a playbook for companies looking to implement AI.
We’ve seen over the past year or two that appetite for artificial intelligence has started to take off, with more and more companies progressing from small scale pilots and experiments into business-as-usual implementations.
AI in customer service and logistics
Two areas seem to be leading the charge: customer service and logistics. As retail has shifted away from the store, customers want their problems resolved online immediately, which has opened up a tranche of opportunities ripe for automation, from chatbots and triage to dispute resolution. Since the early days when chatbot technology was just beginning to be brought in to serve these purposes, many customers have shown annoyance at having to talk to a robot – and there’s still a long way to go to convince customers that by handing them off to a machine, brands aren’t summarily sacrificing the customer experience in pursuit of cutting costs. However, with businesses finding that as many as 60% of queries to their helplines could have been solved if the customer had only read the online FAQs, using AI to catch that low-hanging fruit means staff have been freed up to focus on resolving more complex queries.
In logistics, that same shift to ecommerce has meant that AI has been vital for keeping fulfilment centres firing on all cylinders, and providing delivery scheduling and route optimisation – particularly in the UK, to help drivers make the most of an incredibly old and suboptimal road network.
Fundamental improvements rather than PR buzz
AI is beginning to make inroads in these areas, and the appetite is there across the industry for businesses to implement it. Yet many retailers are feeling they should be doing something with AI without having pinpointed quite what it is, in case they miss the boat. As a result, and as happens so often with hot new technologies in retail – we’ve seen the same happen with the blockchain – a lot of the ‘AI’ executions we’ve seen are still just leveraging the buzz carried by the term without using it to make any fundamental improvements to the business.
One execution that has really impressed, however, is Walmart’s Intelligent Retail Lab (IRL), the grocery giant’s test store for AI and other new technologies in Levittown, New York. Rather than serving merely as a testbed for Walmart, the store is being used to open up a broader conversation with both customers and associates on topics like big tech, data privacy, and surveillance – and to educate them about AI and why it’s being used. Touchscreens and interactive wall displays help to explain the function of the store’s AI-enabled cameras and sensors, which monitor stock levels and automatically schedule restocking by comparing what’s on the shelf to predicted demand. In IRL, the theme of “transparency in tech” is both metaphorical and physical: the entire back wall is devoted to a glass screen that gives customers a clear view of the servers that make up the store’s built-in data centre.
The store plays a key role for Walmart in having two important conversations: allaying consumer concerns with regard to privacy, and helping its associates understand that Walmart intends to use the technology to improve their roles rather than render them obsolete. The store employs 100 associates, and argues that AI in the store frees them up from menial tasks like stock checking, instead allowing them to focus on customer service, being more creative, and ultimately, enjoying their jobs more.
Where’s AI heading?
Automation is already biting. The conversation that Walmart is having with its staff at its IRL store is an important one, because job losses are already happening, and it’s likely the trend of lower head counts in physical stores will continue. One of the challenges for retailers and brands will be to ensure that the roles that remain are suited to those changing processes and that they make the most of their talent.
Ultimately, artificial intelligence alone isn’t going to be the factor that makes or breaks a retailer over the next five years. Despite the hype, it likely won’t be the difference between a company thriving or going into administration, but for those retailers that recognise the technology as a business driver and a transformational force – and not just an IT issue – AI has much to offer.
GDR’s playbook for AI integration
For those companies pondering whether AI can work for them, here is our playbook on what you should be thinking about:
- 1) Start with the business need: So what’s the best way of bringing AI into your business? Our main message, as with all shiny new technology, is that companies should always start with the business need. There is a risk of unwittingly focusing on “tech for tech’s sake”; finding a problem and then asking if AI can help solve it is far more likely to have a genuinely transformational effect on a business than starting with AI and trying to find something for it to do. This means that sometimes, of course, the best solution isn’t technological and can be found in straightforward analogue improvements to processes.
- 2) Due diligence: Once a problem has been identified and AI looks like the right solution, the next steps are due diligence: finding the right vendor; making sure the systems and data are in place; and ensuring your customers and your staff are ready. Appointing an AI expert within the business, or a neutral third party evaluation, can help ensure the system will really do what the vendor says it will do.
- 3) Test and learn: With everything in place, now you need to test and learn. Tests should be made on a small scale to allow failure and rapid feedback loops, but at the same time, test across a wide range of use cases. What might work in a retailer’s “best” locations might not go as smoothly once rolled out across the estate. Test stores, and Walmart IRL is included in this, tend to be placed in areas where customers (and staff) are going to be the most receptive to these changes – but for a true picture, underperforming stores, those in less affluent areas, and those with an especially high (or low) footfall need to be included in the trial. Similarly, beta trials and early access can provide a real advantage to retailers trialling AI in customer service by leveraging the most engaged and forgiving customers, but again that alone won’t give a diverse representation of the customer base in its totality.
If you’d like to talk to us about how your brand could learn from these principles and use them to improve your logistics or service offer, this is exactly the type of project we work with our clients on. Drop an email to Kate Ancketill at email@example.com.