Photo: Groceryshop
Is AI at the center of the retail technology universe?
Based on my non-scientific observations of the industry over the past 30+ years, the average lifespan of an innovation fad is about five years. Big Data was big for about five years earlier this decade, as was RFID last decade and continuous replenishment systems the decade before. There are, however, some developments that are clear long-term trends, including e-commerce, mobile computing and, it’s becoming clearer every year, artificial intelligence (AI).
Again, based on my observations, more than half the booths at this year’s Groceryshop in Las Vegas contained some element of AI, either in their signage or handout materials. Several sessions were dedicated to the topic and Coresight Research released a survey on AI in retail that included these insights:
- Real personalization is challenging for retailers.
- Forty percent of retailers plan to invest in AI/machine learning solutions in the next 12 months.
- Retail is leading the spending and adoption of AI. Healthy venture capital funding for AI startups is paving the way for a robust future industry.
- Most retailers believe that AI will help solve the problem of waste.
- Cost and accuracy are the two main factors that cause retailers to limit the scope of AI deployment.
AI is now going to be used for just about everything — certainly, for optimizing critical functions like pricing and promotions, but also for supplier management, logistics (last mile) and even in-store activities like task management. Imagine using AI to help automate the restroom cleaning process.
As Farhan Siddiqi, the chief digital officer at Ahold Delhaize, said at the conference, “Human/computer interface, automation/robotization and artificial intelligence — put all three together, and that’s where the magic happens.”
Discussion Questions
DISCUSSION QUESTIONS: What does the near-term future of artificial intelligence in retail look like? What will be the most broadly deployed AI retail application in the next few years? What comes after AI?
Poll
BrainTrust
Joan Treistman
President, The Treistman Group LLC
Peter Fader
Professor of Marketing, The Wharton School of the Univ. of Pennsylvania
There is no doubt that retailers will continue to invest heavily in AI. Tech companies are telling retailers they have to, and because their competitors are doing it they feel they must as well. The problem is twofold: it’s expensive, and there is never a guarantee that what you invest in is going to pay off. That makes any new venture a challenge, especially when it is something unproven and without much knowledge on how the customer will interact with it.
Backroom AI or any behind the scenes technology is safe, and it allows a company to iron out any difficulties without the customer fully knowing what is taking place. However front-end technology, i. e. robots and other technology tools can be more challenging, especially when the customer is forced to use something he or she may not want to. So caution should be employed here, but because it’s a race to see who will have the latest and greatest technology first, I see many retailers missing out on other more important initiatives that will boost sales and ultimately build a more significant business.
Don’t get me wrong. I love technology, and I love being the first to have the latest device or software. I choose to do that, but when customers are forced to use something they are not comfortable with, chances are they won’t be that retailer’s customers for long.
AI has tremendous potential and it has already been proven in many areas. As in any area of utilizing a new technology that has great promise, think Big Data, there must be a solid and well thought out plan specifically focused on the problem you are trying to solve. It cannot be everything at once as some senior executives will treat it when a shiny new object arrives.
Too many times, they want to use it but the strategic and implementation plan is not well thought out. I am confident in the benefits of AI and do not want to see it go the way of many other technological trends that had great promise and were simply not planned or executed appropriately.
I like the idea that AI can reduce waste, whether it’s fresh tomatoes or cashmere sweaters. A walk through most mall apparel retailers during any week of the year will find anywhere from a couple of rounders to an entire room of prior season inventory. Every week, all year long. It’s particularly bad in January and February. It seems overbuying is a common malady. Maybe AI can help inform the retailer better as to breadth and depth of the buy. I hope the “what” is still very much driven by real human intelligence, aided by AI. Apparel retailing is still a sensory and emotional activity. I hope that means humans have a role for a long time to come.
Retailers are investing in AI to streamline operations. Since it’s all about software I think the benefits will be visible soon, if there are impactful benefits. Machine learning aggregates data and displays it in a meaningful, i.e. usable form. If the right team is in place retailers can leverage the potential cost and time efficiencies of AI.
“Is AI at the center of the retail technology universe?” I sure hope the answer is no!
Meaningful insights about buyers and buying patterns should be at the center (and all too often, this isn’t the case). But that’s different from the tools that enable them.
It’s just a rebirth of the same old 1990s fuss (and, ultimately, frustration) about CRM — it was going to change everything, but the focus on the tools ended up blocking out the focus on their purpose. So count me as a pessimist here: the same thing will happen with AI.
AI is definitely the bright shiny object of the week.
Where it’s most interesting to me is in forecasting. It can bring in non-transactional data to hone the forecast.
Let’s not get too crazy here. A lot of what passes for AI just isn’t.
Paula, I agree 100 percent that forecasting, especially in the grocery segment in efforts to reduce waste, optimize replenishment, fulfillment for online orders, and even assortment is where the potential is high for significant impact.
Merchandise planning in other segments is a good target as well. It’s a way better way to get customer information into the plan than creating attributes or alternate hierarchies or, God forbid, a fourth planning dimension.
The few places I’ve seen AI truly successful, however, is in demand forecasting (spot on @Paula @Ricardo) and online personalization. These are predictive arenas, where large amounts of data can be collected and consumed for future action, less real time.
My crystal ball tells me that AI will become more embedded. It will be invisible and an inherent part of solutions and planning over time, not something you add in separately, making it all the more difficult to discern what’s really AI and what’s not (as if that’s not hard enough already).
It’s too early to say whether AI is the tech flavor-of-the-month for retail, or if it’s here to stay. The challenge – as always – is to stay focused on customer needs, then find a solution — which may or may not be AI. But the current trend has been to take the AI hammer and randomly whack everything with it. AI can be an amazing tool, and there is tremendous potential for it in retail, but it’s not a magic bullet.
It’s just not that simple to consider AI at the center of the retail technology universe yet. There are far more fundamental challenges for retailers and brands to prioritize and consider before making the evolutionary leap to an AI-led strategy and operations. Yes, retail is all about the delicate balance of art and science, and AI has started to be slowly integrated the mindsets and culture of retail organizations. However by simply leveraging AI capabilities, companies could initiate things by taking a crawl/walk/run piloted approach. Becoming a more prescriptive, customer insights-led organization involves parallel organizational, cultural and technological transformations.
As opposed to specific technologies designed to provide targeted functions, like RFID, CRP, etc., Artificial Intelligence is more of an evolution in computing — one that takes Big Data to the next level, so to speak. AI is a type of technology that includes things like machine learning, and is being used in virtually every aspect of business, across industries. Yes, AI will be around for a while. Another buzzword, “quantum,” which is a revolutionary computing design, is also already taking a hold in industries around the world and I see it also being adopted by retail in the not-too-distant future.
If you believe the vendors, then AI is absolutely the best thing since sliced bread! Is the reality just as pretty a picture? The potential is definitely there. At GroceryShop, Ocado gave great examples of the potential to improve forecasting. Their claim of reducing waste to 0.7 percent raised quite a few eyebrows at the show. Perhaps the biggest challenge retailers face is not where to apply AI, but who will do it? Finding the right talent, people with the right background and skills to apply AI, may be the #1 thing that holds retailers back. This labor is in high demand, not just in retail, and retailers will need to be willing to not only pay them competitively but ensure they are coming into a corporate culture that will allow them to thrive.
In marketing research, some have claimed you can’t do it without AI – and of course that’s a load of crap. In retailing, we are starting to hear the same thing and to Peter Fader’s point, it’s still a load. There are some uses of AI (not to be confused, as it often is, with automation) that can make a business more profitable – sometimes it can produce better predictive models and I’ve seen it work well in logistics. A panacea it is not, however. You need smart people to understand what goes in, what goes on, and what comes out.
AI in retail in the near-term looks like solutions that try to extrapolate actionable insights to drive sales from masses of previously indecipherable data, easing the considerable burden on retail IT, ops, and merchant teams. It will be deployed more broadly if solutions result in clear reduction of waste and/or revenue gains, and happily and seamlessly partner with human retail workers. Customer-facing innovations are less likely to take hold than their backroom counterparts.
I believe that the future of AI in retail is wide open. We have the incredible opportunity to leverage real-time cloud connections that will allow retailers to deploy a central nervous system approach to retail. We can then sense and respond to situations in real time just as our bodies do by creating networked visibility/connectivity to 1,000’s of locations in real time, identify patterns and automate solutions. When all retail technology moves to the cloud and the silos of separate systems for eCommerce, store POS, inventory and labor management disappear, the true value of AI will be realized. The ultimate synergy of people, process and technology can be realized.
AI is the most recent advance in analytics — retailers have been on a decades-long journey to gain insights from data. Making better decisions (over time and in real-time) across the organization is essential to connecting with customers, employees and suppliers.