Artificial Intelligence (AI) in Retail

Reading time: 4 minutes

Introduction

Scan the retail press and various eCommerce tech websites and you find no shortage of tech businesses claiming to use AI as a facilitating technology within their products and solutions for retailers.

It’s not confined to retail.

Finance, marketing and security are just some of the sectors where AI has been used to transform markets. And they are way ahead of the game compared to retailers.

What is AI?

In its most simple form AI in computers takes decisions without user interaction.

So is conditional formatting in a spreadsheet AI? Well, sort of. Whereas most software follows rules: “If this, do that, then do this otherwise stop,” what makes AI distinct is its ability to learn and apply that learning.

Machine learning, as it is known, is often used interchangeably with AI. It is a key component in true AI applications. Don’t be fooled by the vendors who talk of AI but not of machine learning.

AI will follow rules but circumvent them if a better outcome, ideally based on your goals, can be achieved from applying learning.

AI’s biggest benefit is its ability to crunch through millions of rows of data from multiple sources, draw conclusions, apply learning and take action.

The fear that the machines will take over is based largely on dystopian films. The truth is AI does the heavy lifting.

AI should provide high-level automation with all the necessary interfaces for manual intervention.  It should be unified and developed specifically for the problem in question with objective-driven optimisation and predictive simulation.

Crucially though it will learn and scale to maintain relevance and therefore performance over time.

AI in retail

The challenge faced by retailers interested in introducing AI into their organisations is there are so many different aspects that can benefit:

  •      Warehousing
  •      Supply chain
  •      Payments
  •      Fraud prevention
  •      Marketing
  •      Customer service
  •      eCommerce merchandising and product exposure.

The solution is not to have an “AI strategy” in the hope of introducing it across disciplines, but to ask what problem you are trying to solve or what business strategy you are trying to support.

The next question is “can AI help with this?”

In the case of online merchandising where there may be thousands of products, mountains of historical transaction data and need to increase profitability over conversion rate, or vice versa, the answer is yes.

A product like Apptus eSales when pitched in an A/B test against manual merchandisers, made 101 sales compared to 1 sale from a manual approach.

How to prioritise AI investment?

Most businesses are fundamentally trying to achieve simple, but not always easy, goals: to sell more, to be more profitable and to reduce friction for customers and employees.

If the goal is to sell more profitably, then focussing AI at the front-end i.e. merchandising and reducing customer friction should inform the investment priority.

Conclusion

AI offers the possibility of improving operations to a level that could accelerate growth for some retailers in ways that could create huge, potentially unbridgeable gaps between them and the rest of the market.

AI certainly is a big part of the future – the question is when that future will start in earnest?