Making a case for AI in retail applications

by Vawn Himmelsbach, a technology and business writer
Montreal, Quebec - September 29, 2021

Maybe the C-suite is wondering how to increase productivity in a post-pandemic world. Maybe they’re looking for supply chain efficiencies, or new customers and new markets. IT directors in retail know AI can help the business gain insights from the data it collects. But they also know AI isn’t activated by pressing a button. It’s a long-term investment — not just in technology, but in human capital.

While AI will impact almost every sector, it will be a competitive differentiator in retail. Retailers are already using chatbots, recommendation engines and price personalization platforms to find new customers and keep existing customers happy.

AI is also being used to help customers “try on” products virtually through augmented reality, and investments are on the rise in robotics, virtual assistants and inventory management, according to the Retail Council of Canada

While AI will be a competitive differentiator for retailers, it takes time to deploy and start seeing the benefits. So how do IT directors convince higher-ups that they need to invest in AI now, rather than take a wait-and-see approach?

“The choice to embark on the AI journey needs to come from the top. It’s a strategic direction that you take, not a project that you decide to fund, because it requires significant investments at multiple stages and in multiple places in the company.”

Clement Bourgogne, VP Strategic Programs at Scale AI, Canada’s artificial intelligence supercluster.

But this is where IT directors can have a tremendous impact in helping the business understand what problems could be solved with AI, he says, and how it could be applied within a specific division or business unit. “They have a very critical role in adoption, but also in deployment — because if you want a tool to be successful, it needs to be based on a business need and it needs to drive business value.”

In retail, for example, a popular application is demand forecasting: trying to predict how much product you’ll need at a given point or during a given period. Historically these predictions have been based on sales figures from previous years, which in many cases are no longer relevant in a pandemic or post-pandemic world.

“It’s not very accurate or it’s not based on a real, tangible trend. It’s just saying, ‘well, what happened in the past will happen again,’ and that’s where an application like AI is extremely powerful because it’s able to correlate what happened with certain events and provide a better, more accurate prediction for the year to come based on actual data. It will also be able to do it at a much more granular level.”

Clement Bourgogne, VP Strategic Programs at Scale AI, Canada’s artificial intelligence supercluster.

Another use case for AI is inventory management, for both business-to-business (B2B) and business-to-consumer (B2C) companies. While this is closely linked to demand forecasting, it relates specifically to how much inventory you should have on hand — a critical issue for retailers right now during supply chain disruptions and shortages. 

“Inventory costs a lot of money in terms of working capital, so if you can optimize that, it can help drive a lot of savings,” says Bourgogne.

In-house IT departments will typically collaborate with service providers that are experts in AI. But there’s often reluctance to deal with less-established players.

“Usually when companies approach us and say, ‘I’d like to find an AI service provider, can you help me find someone who’s done demand forecasting for the last five years in retail and can start a project next week?’ The reality is this does not exist.”

Clement Bourgogne, VP Strategic Programs at Scale AI, Canada’s artificial intelligence supercluster.

“There’s no long list of companies that have been doing this for five years. Everyone is just starting to do it, so you have to be willing to partner with less-established organizations —because that’s where the talent is.”

Collaborating with the right service providers allows you to start small and fail fast, he says. Most AI projects don’t fail because of the technology; they fail because it’s not integrated into the company’s processes and workflows. In other words, it’s an adoption issue.

For example, say a retailer decides to start using AI-based dynamic pricing (where prices fluctuate depending on various factors such as the weather or proximity to an event). Franchisees are then told to increase or lower their prices whenever the algorithm tells them to — but without training, they might be hesitant to trust the algorithm, especially if their competitors aren’t changing their prices.

“The algorithm might be right, but ultimately if the franchisee doesn’t follow the recommendation, your technology is worthless, so that’s why change management and helping people within the organization is so critical.”

Clement Bourgogne, VP Strategic Programs at Scale AI, Canada’s artificial intelligence supercluster.

AI is well suited to solve certain business problems, but not all. “The starting point is to educate, not the technical resources,” says Bourgogne. “It’s not about how do you launch your project, but figuring out where you can deploy it strategically inside the organization — and it requires training of the business people, the managers and the executives of the organization to understand the potential applications of AI.”


Made possible through the
financial support of
Gouvernement du Québec
Gouvernement du Canada