Changing the game: reimagining the retail industry powered by AI

But retailers are facing intense headwinds. The cost of products, manufacturing and transportation are all rising; suppliers are often unable to meet delivery timelines; demand from customers is unpredictable; new channels such as e-commerce need to be managed; the imperative to reduce environmental impacts cannot be ignored. And all of this is made that much more difficult with outdated technology.

It’d be foolish to suggest there is a silver bullet or panacea for these woes. But the context we’re operating in does provide some clues for what to do. We’re on the precipice of a revolution, not unlike the dawn of e-commerce in the mid 90s or the launch of the iPhone in 2007: artificial intelligence (AI) is transforming the world around us. Every decision a retailer makes can, and will, be informed by AI — resulting in unprecedented leaps in operational efficiency and improvements in customer experience.

Retailers would be wise to consider how they can use AI to build competitive advantage. Those that don’t invest today risk the same fate as those that ignored the e-commerce trend: they’ll disappear into the retail history books. So, where do they start? Broadly speaking, there are three main themes to adopting AI in retail…

1. Interconnected data as your crystal ball

If only it were possible to know exactly what consumers will want in the future, down to the details of how many, what color, in what stores and what price they’d be willing to pay. The next best thing to having a crystal ball is using your data to manage uncertainty.

Retailers have always been fueled by data, going back to the early days of pen and paper recordkeeping or Lotus 1-2-3 (I’m aware I’m really dating myself here!) The modern AI-enabled retailer must bring together data from different sources to create a single version of the truth that enables data-driven decision making. Supply chain, demand, pricing and customer data can all be combined into data products or a data fabric that allows business applications and end users to access data and insights as needed.

Bringing together data in this way will allow a retailer to create a demand forecast powered by machine learning that becomes the beating heart of the business. With a data-driven and continuously improving consensus on what consumers are most likely to buy, when and how, retail teams can plan their buying, merchandising, allocation, pricing and promotions more effectively.

For example, with a more accurate and more connected demand forecast, the planning team for a major global sportswear retailer can allocate individual SKUs to exactly the right stores on a daily basis, resulting in higher levels of sell-through and increased profits.

2. Probabilistic decision making

This sportswear allocation example illustrates how building AI that leverages a variety of data sources has the potential to disrupt entrenched business concepts, notably “business rules.” Every retailer has these rules and uses them to make decisions. But the truth is that rules alone were never meant to be the basis for good decision making. The constraints of technology have made retail decision making models what they are today.

We’ve ended up with a lot of rules-based systems, not because they work best, but because that’s all we could handle. Legacy technology systems and the rules they brought have not prioritized innovation and creativity, but rather standardization and compliance. Merchandising, buying, planning and pricing have all been slaves to the rules imposed on them by a sub-optimal approach to technology.

But things don’t have to be that way anymore!

The right way of doing things for a retailer no longer means a single right way to the exclusion of other options. Data-driven decision making is probabilistic by nature, with infinite options allowing for more creativity, flexibility and innovation. 

For example, an AI-powered department store could send different marketing messages to each one of its millions of individual customers, or set a different pricing strategy for each one of its thousands of SKUs. The days of planning a range by taking last year’s sales and arbitrarily adding a growth assumption can be left in the past. AI gives you the flexibility to adjust plans quickly and easily. The end result is increased productivity, happier customers and stronger margins.

3. Sustainability

In addition to the bottom line benefits of leveraging AI, there can be enormous environmental benefits. Using AI to optimize buying means raw material use is limited to what you know will sell. Machine learning models can optimize stock movements to and between distribution centers, reducing thousands of miles of unnecessary emissions. And utilization models can help drive down energy, transportation and other resource uses.

Getting started with AI in retail

How does AI fit into a retailer’s existing technology ecosystem? Firstly, we have to consider the state of the tech that they’re currently using today. The systems that run retail businesses may include a myriad of legacy operational and planning systems, plus systems for accounting, HR and much more. On top of this, many retailers will have a patchwork of point solutions for planning and analytics.

To leverage the full transformational potential of AI, retailers need to build a new layer of tech: an AI layer that sits across the value chain and across their existing systems and infrastructure. The ideal technology architecture to do this is a composable one with individual tools and systems working together. This will offer retailers greater flexibility, and create opportunities to infuse AI across the entire retail value chain.

In conclusion, as the tides of uncertainty continue to wash over retailers, the time is now to invest in reimagining retail powered by AI.

For more information on the game-changing potential of AI for retail, download our latest retail whitepaper below.