Give AI a chance

Give AI a chance

Retail

STOrai Magazine

STOrai Magazine

260 week ago — 7 min read

Background: Former CERN scientist and Founder of Blue Yonder, Michael Feindt speaks to Shiv Joshi about the applications of Artificial Intelligence and Machine Learning in retail, today and tomorrow.

 

According to a Markets and Markets report titled ‘Artificial Intelligence in Retail Market’, the global market for AI in retail will be $5034 million by 2022. Such is their influence that Artificial Intelligence (AI) and Machine Learning (ML) dominate every meaningful conversation about business and technology today. What seemed like science fiction a few years ago are increasingly being used in the real world to solve problems and improve efficiencies with remarkable benefits, thanks to scientists like Professor Dr. Michael Feindt who are working to bring their benefits to retailers across the world.

A former particle physicist at CERN, Dr Feindt informs that AI and ML are in fact technologies that have been in existence since decades. He built his first neural networks 20 years back. It is the potential that they presented in solving complex problems that inspired him to focus on applying them to retail and starting his own company Blue Yonder.

In an insightful interview, Dr. Michael Feindt answers the real question about AI weighing on the collective minds of the industry despite the optimism: ‘Does AI have a real world impact on business?’

How can AI and Machine Learning help retail?

To put it in simple words, AI and ML affect every area where decisions have to be taken. In retail, they have been used primarily in supply chain, for replenishment. This is an area on which we are working with most of our customers. We are trying to individualize and optimize decisions that are taken in a retail store daily. These have usually been done by humans or by relatively simple computer programs. AI can help answer questions like how much of every item to order such that there’s enough on the shelves but not too much especially for fresh food, perishables and time-bound products like seasonal fashion.

We have many large bricks-and-mortar retail chains as customers. For them, the effect of AI is significant as the items have to distributed to all the stores. One cannot have empty shelves because it results not only in loss of sales but also the loyalty of customers. It is difficult for humans to take accurate replenishment decisions for 500 or a 1000 stores due to many uncertain factors involved. That is really the specialty of our prediction software —it can handle uncertainty. No one can predict the exact number of products that’ll be sold on a given day, but AI doesn’t just churn numbers; it also builds a risk profile.

 

AI and ML are also extensively used in pricing. Variables like the location of the store, weather, type of product, competitor’s prices are used to determine the ideal price of a product. This is important in competitive markets where you need to hit the right spot with utter finesse to make sure you win market-share as well as make profits. And this is for both online and offline retailers.

 

Customer service and interaction are also domains that are greatly influenced by AI. Automated stores and autonomous checkouts are examples of the same.

 

How is AI superior to human intelligence?

I think AI surpasses humans when it comes to decision making and predictions. There are so many factors that influence sales now like the weather, which day of the week it is, the competitor’s price, holidays and so much more. To take into account all of this is very complicated and only good AI algorithms can learn from these observations and help make precise decisions. AI learns from historical data that a retailer has. It also learns from experience and modifies itself mathematically to identify subtler changes and takes the best decisions based on these changes… and all this is done automatically. It learns every day and self-optimizes. We don’t stand a chance against such a system. Industry leaders like Amazon are relying heavily on similar systems for prediction and complete automation.

 

Human decisions have sub-conscious biases. For example, retailers always buy too much during Christmas because they are afraid that it will not be enough. This leads to them having a surplus which they then deal with by giving away freebies and promotional offers. A machine is unbiased and fearless; it would never make this mistake.

 

The overall strategy and policy will still be steered by humans. The machine ensures that the strategy and policy guidelines are met by every single store, every single day and helps optimize the store in a way that makes the business more profitable. And while the machine is looking after the supply chain, the retailers can focus on how to please their customers and make the shopping experience more enriching and enjoyable. At the end of the day, it is the consumer that decides whether or not a retailer will be successful.

 

How is the future going to be for AI in retail?

I am not convinced that all stores will be without personnel. And I don’t think that online will be everything and offline stores will die. Most humans still want to have some interaction with other humans; AI will enhance this interaction even further.

 

I'm completely convinced that more and more decisions will be taken with the help of AI. And that all the retailers will have such systems in the next 10 years. The big ones like Amazon or Alibaba use such methods massively to get their processes straight, which is why they are so good and they can deliver so fast and get good prices. The biggest mistake anyone can do is to ignore this. No matter what your business is, what its scale, or what your channels are, AI and Machine Learning will add value to your business. Not adapting may mean the demise of your business.

 

Any last words for reluctant retailers?

Just give computers (AI and ML) a chance and see what they can do for you and your business.

 

Article source: STOrai Magazine

 

Disclaimer: The views and opinions expressed in this article are those of the author and do not necessarily reflect the views, official policy or position of GlobalLinker.

Comments