As Chief Evangelist at Google, Nicolas Darveau-Garneau (BMath ’92) spent his time with top e-commerce executives, explaining how AI could drive more traffic to their platforms and increase sales. About three years ago, he noticed executives were more interested in a new topic. They wanted AI to improve the customer experience on their websites and apps.

“If you go to Amazon as a customer, and you're logged in, they know who you are,” Darveau-Garneau said. “Your entire experience is completely personalized to you in an extraordinary way. Also extraordinary, if you're not logged in, what you're searching for, what you look at, what you click on is all analyzed in real-time by machine learning. The whole site changes automatically. It's personalized to you in a very, very powerful way.”

This is what CEOs want for their own digital experiences and, according to Darveau-Garneau, customers are coming to expect this level of personalization as well. It’s a turning point for e-commerce.

“It’s easy to see how one-to-one personalization improves the customer experience. As an avid online shopper, I can attest to the frustrations that come with the search functions on certain sites. Not surprisingly, we are more likely to buy from sites where we can find what we’re looking for,” Darveau-Garneau said, adding that increased sales don’t necessarily mean increased profitability.

Personalization vs. profits

For example, if a shopper searches the web for an ergonomic office chair, an AI program can show tailored results ranked by the likelihood that shopper will buy them — all based on his or her previous activity. Those products may not be profitable for the business, however. As a bargain shopper, Darveau-Garneau said he’s likely to buy something marked 50 per cent off which won’t generate much profit. “I might also buy a chair that most people find uncomfortable, increasing the chance that you’ll lose profit when I return it,” he said.

Without AI, site managers get around these issues by writing rules into the backend, ensuring that a new item is shown in recommendations and search results, while others are less visible. But these rules are difficult to manage manually, interrupting the customer’s personalized experience.

Darveau-Garneau describes the tension between personalization and profitability as a Gordian Knot. A human can't untangle all the factors on their own. Given today’s customer expectations and the fierce competition in e-commerce, Darveau-Garneau believes businesses need AI to survive. “The majority of retailers that did not shift from printed flyers to digital marketing did not survive,” he said. “The majority that did not shift from purely store sales to a mix of e-commerce and in-store did not survive. This is the next wave.”

“The majority that did not shift from purely store sales to a mix of e-commerce and in-store did not survive. This is the next wave.”

Nicolas Darveau-Garneau (BMath ’92)

Nicolas Darveau-Garneau

Humans are still needed

For many workers who fear AI could replace humans in e-commerce, this might sound alarming. But Darveau-Garneau is clear that humans will still play an important role in managing e-commerce sites. Instead of writing rules, they will manage and coach the AI, while testing new strategies to improve the customer experience and increase revenue. In the end, the change will be good for business, good for customers and good for workers. 

Darveau-Garneau believes this change will start with thoughtful leadership: “Every single executive at every company should take a pause right now. What parts of your business are about to be massively disrupted by AI to make things better? What is it that your customers are going to love? What can you do that's very different? Be open-minded.”