The common-sense approach to AI

Over the past decade, consumer demand for personalization in retail has changed the way the industry operates. In order to drive loyalty, retailers have to have the right price, the right promotion, and the right products for their customers. And today, savvy retailers are using artificial intelligence to ensure they’ve got what their customers want. 

“AI allows retailers to focus on what the customer wants exponentially more than what's been possible historically,” says Gary Saarenvirta, Founder and CEO of Daisy Intelligence. “What mid-market retailers don’t have in volume-buying clout; they have to make up for with a tailored customer focused value proposition”.

AI is here, and it’s happening. But knowing where to start can be daunting, to say the least. Understanding AI and how to implement it isn’t easy, but with a strategic approach that ignores the hype and instead focuses on the reality of an AI solution, it can be done successfully, setting retailers up for success now and in the future.    

Understanding AI

There is no shortage of confusion about what true AI is, but to truly understand what it is, perhaps it is better to first establish what it is not.

Backwards-looking analytics that provide scenario-based predictive models are not the AI of today, says Peter Leech, Partner and Director of Digital Commerce of The Partnering Group, Inc. “That uncovers historic relationships between unit sales and promotion and uses snapshots to predict future promo plans.” But today’s AI, he notes, is “the computer itself dynamically learning what is the optimal and processing recommendations in real time. And that differentiation in speed and marketplace responsiveness is huge.” 

Today’s true AI predicts without using historical examples. It is autonomous and doesn’t need labeled data. It can evaluate billions of options in a way that humans can’t. It is reinforcement learning that learns through trial and error without intervention, says Saarenvirta. “If you need to have examples from the past to be able to predict the future, that’s statistics, not AI,” he adds. Understanding what AI is and is not is critical to its success within an organization.    

Choosing the right AI in retail

If dramatic improvement in performance is a retailer’s goal, they aren’t going to find it with traditional statistical analysis. To confuse matters further, machine learning is just new terminology for predictive analytics’ old school methods. What’s most important is to deliver verifiable financial value, and retailers need to ask the tough questions to determine the technology and vendors they choose can deliver on the promise of AI says Saarenvirta.

The key question retailers need to be asking about is the verifiable economic value proposition, how AI will grow company sales and profits, and also how it integrates with existing human processes. When those questions are answered, retailers can prioritize the big problems they face, and then select from that list those processes where decision making is highly repetitive, complex, mathematical and occurs in large volumes. These are the types of decisions that really would be best done by a computer with the retail merchants setting the strategy and business rules or guardrails which keep the AI from making decisions not consistent with the retailer’s strategy and brand promise.

AI elevates the human beings to play a more strategic role in overseeing the AI and setting business strategy.  “The types of problems best suited to AI are not easy to spot,” Leech says. “It takes a lot of work and commitment from the executive team to determine what problems they are willing to use AI for and at the same time eliminating the human barriers that are inevitable with technology implementation.

Retailers need to be clear about what tasks really still require a good deal of human creativity because they are strategic and creative versus computing tasks which are quantitative and mathematically complex.” As a lot of those core merchandising tasks tend to rank high, it then becomes a question, Leech says, of “finding a partner who has proven experience applying AI in your industry or a transferable industry and can prove that they’ve been around long enough and they have trained the machine effectively to produce ongoing success.”     

Implementing AI

AI is not a standalone technology. It must integrate with a retailer’s strategy, in-place technologies, and have buy in from all parties involved. But most important, AI is a roadmap for change, and not just for the technology side of a retailer, but for human transformation within the retailer as well.

When implementing AI, retailers, Saarenvirta says, “need to be committed that AI will be core to their decision-making engine.” That means those retailers are going to have to adjust some of their processes, change the role of some of their people. “Retailers have to buy into this future vision that AI is going to evolve and change the role of people.” And the people component, that human element, is critical when undertaking AI implementation.

In parallel to the human piece, Leech says, a typical approach for most retailers implementing AI is to pilot the tool. However, he adds, “retailers need to recognize that for these AI tools to be truly accurate, they need to operate in a near-full live environment.” That means AI needs to be provided with all the data to make the recommendations work.

Retailers should have full data implementation to get a good proof of concept to pilot, and to create a good pilot charter that shows what they’re trying to achieve and what that looks like, Leech explains. “You have to be clear as a company regarding what you are buying and what marks a score or KPI improvement that would confidently trigger the move forward to rollout or would cause you to revert.”

AI for now, and for the future

What once seemed to many as a far-off and far-out solution has become essential to driving growth in retail. AI is here, and while retailers are feeling the first waves of disruption from it, there is still so much more to come.

“It’s been a gradual evolution, not this gigantic leap forward, the way it’s portrayed in the media,” Saarenvirta notes. “AI is a technology that can do some amazing things, because there is more computing today than ever before. Learning mathematical equations is not learning the way human beings learn, or biological animals learn. AI has to be grounded in practical reality, not caught up in hype.” 

At its core, AI tackles problems that people aren’t good at and leaves people to do what they excel at.

Embracing AI establishes a competitive advantage, and not doing so means missing out on countless opportunities. But, it must be done right—without the understanding what is right for a particular retailer’s needs, time and money will inevitably be wasted, and cause irreparable organizational pain. And as retail competition isn’t showing any signs of slowing down, the time is now for retailers to think smart, thing forward, and think AI. 

To learn about the importance of AI for independent retailers, read our blog here. Or, to learn about the use of AI in convenience stores, read our blog here.