Empowering retailers to see into the unknown with AI

by Michael Feindt, Blue Yonder

9 October 2019

I’ve got a question for retailers who have a fear of the unknown: in reality, do they really know everything when it comes to their core operations, from stock levels and price points to distribution volumes and promotions?

Most decisions are still largely made on gut feeling or by an individual who believes their experience will always outperform a computer, which in today’s technological environment is nonsensical.   

There’s a lot of hype surrounding artificial intelligence (AI), and Gartner research states that the number of enterprises employing AI has grown 270% over the past four years; and yet, many companies are still sceptical about what it means or how it could benefit their business. There is a clear purpose, however: retailers can use AI to illuminate areas of the business that were previously shrouded in uncertainty.

When unhindered by human interference or pride, AI and Machine Learning (ML) can clear the fog surrounding the unknown, ensuring that all decisions are made using accurate statistics and data. Retailers need to take as many uncertainties out of the equation as possible so they can predict, act and evolve.

In practice, AI can be used to calculate hundreds of millions of probability curves for customers each day. Item numbers, multiplied by the number of their stores over a three-week period, can be analysed to ensure that money is spent on the most probable extent of stock at any given time. A human couldn’t possibly keep up with this speed of thinking.

The benefit of using these calculations and trusting AI is that you can make decisions in real-time, considering all eventualities in a matter of seconds. Whether it’s a natural catastrophe, consumer trends found on social media or temperamental weather, if the data has been inputted correctly and is consistent, the system, via ML, can learn and react. This way retailers can ensure accuracy with key details such as orders and price points.

Trusting the machine

So why aren’t all retailers already using AI to make decisions? The answer lies in a study by Mirko Kremer and Stefan Minner exploring human bias in inventory decision making. It found that people want to fight against systematic intervention, choosing to rely on personal experience and “gut feeling” to reach important decisions. You must tread incredibly carefully, however, as a poor decision could cause serious damage.

Naturally, most employee’s instinct is to attempt to fix problems themselves, especially if it’s a specialist area that they’ve been tasked with overseeing. However, people need to stop looking at AI and ML as mechanisms that will run them out of jobs. 

Even when this internal battle has been won, for AI to be a success it needs to be implemented in the right way. Retailers need to approach AI with consideration, not just using an external vendor to oversee the organization’s transition but also employing someone internally who has adequate expertise. Having an internal data scientist or statistician makes it possible for technologies to be introduced with greater focus on the company’s values.

It’s vital to allow this data expert to conduct analysis unhindered and give them access to C-Level decision makers, so they can have a direct impact on the business. Currently, there are far too many vague decisions, so having the right personnel in place to make sure that the benefits of AI are truly implemented is a necessity. To really transform the company, a culture change must be driven across the entire organization, ensuring everybody gets on board with AI driving decisions for them. The changes may be drastic and range from changing internal processes to structures, but only this way can retailers truly maximise the benefit of an AI-driven way of doing business and increase profit margins.

Seeing around the corner

Trusting in AI may appear daunting for retailers, but having an implementation strategy in place that follows the above will make retailers more streamlined, cost-efficient, supply chain-friendly and, ultimately, attractive to consumers. This will differentiate the company from competitors in a retail climate where most are still choosing to steer clear of ‘the unknown’ – AI – which is ironically leaving them in the unknown due to all the insight and decision-making power they are missing out on. Ultimately, those who embrace AI will be ahead of the curve, gaining a key advantage in an industry that is only becoming more competitive with each year that passes.


Michael Feindt is founder and chief scientific advisor of Blue Yonder, a JDA Software company. Blue Yonder is a specialist in AI solutions for the retail sector.