For the purpose of this discussion, AI supports by automating processes. Smart automation means AI can “make decisions” as necessary related to tasks, but not major decisions. In its current state, it can quickly deliver us information, suggest options, and create simulations, while still leaving it up to leadership to make ultimate decisions.

There will always be a combination of arts and sciences in the retail landscape. The clear advantage of leveraging the power of artificial intelligence and machine learning is that retail businesses can now make decisions backed by data intelligence. Tasks that were once mundane and repetitive could be automated, with a more prescriptive approach powered by AI.

However there has to be a balance in our society. There is a delicate balance that the retail leadership teams have to ensure, where intuition and experience have an influence on strategic decisions as well. Veering too far into fully automated decision making algorithm is too extreme, as there has to be a flexible model for executive teams to account for.

Just as the artificial intelligence machines are continuously learning and adapting, we as a society are on a parallel path in terms of how to leverage and maximize the benefits of AI without removing the human element.

Stories of people who got lost or worse while blindly following a GPS are common. The use of AI will put many companies at similar risk if the company lacks strong leadership and a clear sense of what AI is and is not. No algorithm can mimic empathy.

Business leaders need to understand the “art of the possible” with AI. That means they also need to understand what the limitations are.

Using AI in operational decisions doesn’t mean the company has an analytics culture. We use AI every day – a Tesla driver is a good analogy. Driving a fully autonomous Tesla doesn’t make the driver any more analytics-oriented than someone driving a regular car.

Laggards in retail will be passive and will benefit when the use of AI becomes pervasive across all applications. Leaders will truly understand what AI can/cannot do, and make sure to take advantage of AI where it can make a difference.

First and foremost, they have to trust the numbers. In my work, people like to use AI as a buzzword but the beneficiaries simply don’t trust the science. They still believe that 30 years of experience is more accurate. It’s not. The proper use of the appropriate data will help you run your business more profitably. Dashboards don’t help your business – using the insights in a meaningful way will.

The point here is that AI has a great deal of potential, but can only be deployed with confidence when it is smart enough to understand when its recommendations are nearly certain, AND when the algorithms are not certain enough and decisions should be passed to humans to intervene. Deploying AI is a long-term journey and any leaders involved in its deployment need to be learning about capabilities while remaining careful about what its deployment will mean for employees and customers. For some businesses, 90 percent or 95 percent accuracy will be of sufficient benefit to risk alienating people with the 5 percent or 10 percent of mistakes (or awkward recommendations). For other businesses, deployment should wait until accuracy reaches much higher levels. Similarly, I think there is a huge opportunity for some brands to define their customer service by remaining human while competitors and substitute brands deploy AI – and potentially miss the mark. Competing is about delivering value and differentiation. Dealing with AI is no different than many other technologies that have come,… Read more »

I recommend reading the referenced Knowledge@Wharton interview of David De Cremer. I have ordered his book. In the interview he refers to Moravec’s paradox: What is easy for humans is difficult for AI and what is difficult for humans seems rather easy for AI. In today’s business world there is no doubt that AI is being used to model repetitive behavior and it does a pretty good job in those circumstances. To move to the next evolution, in which AI operates within a social context, leaders must fully understand AI’s strengths and weaknesses and invest in the necessary soft skills that mirror and complement AI’s hard skills.

Retail leadership will have to ensure the right people are hired and internal expertise is developed. They can’t think of this the way they started off with e-commerce, using out-of-house capabilities to the detriment of learning and building strong teams.

Necessary skill sets will be non-linear thinkers, people with cross disciplined backgrounds, technical folks and there will also have to be team structures that allow for collaboration, not silos.

Too many of us think of AI as a task-oriented technology. It is not that today and in the future it will be very far from it. If we compare analytical thinking versus systems thinking, AI is systems thinking. It doesn’t drill down, it steps back and sees how possibly unrelated information actually becomes a part of the final action.

Suresh’s example of the autonomous vehicle is one that I was going to use. I was in a Tesla this weekend and, though not self driving, it certainly was “self-thinking.” Another great example is the work done UNC relative to diagnosis. AI absorbs over 200 times more information than a human and is almost 100 percent better at diagnosing healthcare issues.

The key for leadership is to ask “how do we execute the information?” rather than “how did we get it?” And for the future don’t imagine that development in AI is a straight line. The competence of AI will be well beyond our imagination.

In today’s increasingly complex and dynamic retail landscape, AI is no longer optional but required to help retailers make the right data-driven business decisions and allow their strategies to come to life. The real challenge facing most retail leaders is developing the right balance and processes to use AI capabilities properly to maximize its true potential in alignment with enterprise objectives. Developing a culture of AI-informed practices which consider key qualitative insights isn’t easy, but the best retailers should be aiming to achieve this reality.

There are two fundamental chasms for retailers and consumer-facing industries as exemplars of economic sectors collecting unprecedented amounts of data. One is the actual awareness of the inherent potential and insights residing in the data collected – this cohort is steadily decreasing. The other, and more severe gulf, is having the organizational mindset and culture to create customer value from that data. Those are prerequisites before we can entertain the potential and leverage that artificial intelligence, which is much broader than machine intelligence, offers in profoundly enhancing managerial decision-making. Consider that spreadsheets as in Lotus 1-2-3 and Excel were the killer apps that launched the popularity of the IBM PC and Microsoft Office, respectively. I suggest that today, AI is the next killer app for the 2020s compute environment – be it at the personal or enterprise levels. What remains elusive is the nature of the platform that AI will spawn and from which mass adoption will occur. It begs a radical construct which proposes that application logics flow through purpose-built AI machines instead of… Read more »

What AI represents to the retail business is a function of its challenges and the business outcomes potentially improved through intelligent automation. It is as simple as that to begin with, and so you need to land on the use cases most material for your business. So it’s helpful to know what an AI use case looks like. Beyond that, there are considerations for prioritizing use cases, governance, ethics, oversight, technology products, technology architecture, data science, training, education, and re-skilling. It sounds big and scary, which is why I think so few retailers can articulate a vision for AI in their business. The easiest path I think is to view AI as an extension of the long-standing analytics function in your business and build out from there with an eye on maximizing the value of investments across technology, people and processes. I would not fixate on building models and deploying them versus purchasing tech products or services off the shelf. Instead, I would view these as part of one “whole” approach to leveraging data and… Read more »

I think we are finally entering the productive phase of the conversation around AI in retail. That is, it is advancing beyond a marketing label applied to business solutions toward an understanding that it is a new essential resource for automating certain processes and enabling decision makers to take more precise, effective actions. (Thank you Dr. George, for reminding us about Moravec’s paradox.) Most routine, repetitive actions will certainly be handled better by trusted AI systems — interpreting shelf images to track inventory, computer automated reordering, price optimization, localized assortments and personalized promotions come to mind. But decision-support may be the far greater opportunity. The AI has an advantage in that it can discover unanticipated patterns by digesting multiple data sources, model the data, and compare new data against existing models to prescribe next actions. Human decision makers who have been freed from managing routine processes and who may not possess data science skills, are more able to apply their judgment to a reliable set of facts and prescriptive alternatives that the AI provides. To… Read more »

It’s the soft “squishy” skills that are a big part of what defines a great business leader, not tech-savvy skills. While AI may be destined to help make difficult, data-driven, carefully analyzed decisions, ultimately, there will always be major business decisions that can make or break a company that need to be made by business leaders. Sometimes even AI can’t replace good experience when business leaders make tough decisions. Look no further than this year’s pandemic and retail supply chains to see how AI would have directed businesses to alter their supply orders and inventory management. Many retailers would have made foolish decisions. Will AI continue to grow and develop to the point where those mistakes won’t happen? Most likely, yes, but there always needs to be a balance in any organization with human thought leading the final decisions. As others have mentioned, this is similar to people blindly following a GPS direction that leads them the wrong way due to a mistake. The key is to understand when it’s best to rely on AI-generated… Read more »

We must be careful not to overestimate the current functionality of AI. Algorithms are immensely helpful for tech leaders to understand their customers, and these elements that help companies design for real customer needs will continue to grow in popularity and function. But leaders will always need to look at the big picture, creating outcome-focused visions for their teams and customers throughout the process, something we call “The Product Mindset.” Only human leaders can truly master the creative thinking and vision-setting that must govern the digital product development process.