MONTREAL — With concerns about artificial intelligence’s cost and usefulness creeping into the hype, executives at Montreal’s All In Conference chose to focus on how they’re using it for real things today. Leaders from large firms went beyond talking about the technology’s vast potential to transform business, opting for more practical stories. Here’s what you need to know.

Just for you: Loblaw has long personalized its grocery offers and recommendations based on shoppers’ buying habits, but could previously only show a few variations, said Lauren Steinberg, senior vice-president of media, loyalty and digital, on stage Wednesday. 

And customers’ preferences “are so vast,” Steinberg said, adding thatLoblaw’s ability to personalize has greatly expanded with the help of AI.

Take last year’s holiday season. The retailer used generative AI to produce thousands of versions of its homepage, displaying different themes and products depending on the visitor. 

Shoppers whom Loblaw’s data identified as baking enthusiasts might see a photo of a kitchen messy with flour and cookie-cutters, with a feed of ingredients and equipment they might want to buy below. Customers who don’t like to get their hands in the cake mix, but who love to host, would instead see readymade brownies. And budget-conscious buyers would see sale items.

As a result, shoppers were more likely to return to the site, and spent $3 more per order—a lot of extra revenue in the aggregate.

Loblaw could manage all those different pages because it was using generative tools. Staff employed marketing software Jasper to produce the copy, and Dall-E to make the images. It’s now using the “personalization engine” for all seasons. “Our merchants want to show you the best turkey price during Thanksgiving,” Steinberg said. “If you don’t eat meat, you shouldn’t see that turkey promo. Historically, you have; you no longer will.” 

Step through: Corporate adoption of generative AI remains a work in progress. With some staff turning to ChatGPT on their own, many large firms are setting up IT department-approved workspaces where employees can try large language models (LLMs) from the providers their firm has chosen. Wealthsimple, for example, offers employees a mix of open-source offerings like Meta’s Llama and commercial ones like OpenAI’s ChatGPT or Cohere’s Command R.

Corporate giants like CIBC, Loblaw and Telus also have these workspaces. “We built what I call a ‘walled garden,’” said Pamela Snively, the telecom firm’s chief data and trust officer. “It was a safe place to play for our team members.” At Telus, staff built a customer-service chatbot. Employees at other firms are using their AI workspaces for routine tasks like writing emails and summarizing documents.