A big challenge for retailers is increasingly how to use artificial intelligence to make their businesses more efficient and to better convert potential sales.
Some AI strategies make shopping fun for the consumer.
Vancouver eyewear seller KITS Eyecare Ltd. (TSX:KITS), for example, has augmented reality that uses AI to enable shoppers browsing its website to try on frames and see themselves wearing the eyewear.
Other AI uses focus on operations: Customer service, inventory management, store selection and more.
One problem, retailers have told BIV, is that they believe they have insufficient data for AI investments to make sense.
Retail consultants urge those merchants to get started at collecting data so they can eventually put it to use.
“You need lots and lots of data points,” said DIG360 retail analyst and principal David Gray.
“If a small operator is trying to use AI using internal data, they may not have enough data to really do it.”
Paul Dragan, who owns the two-location Reckless Bikes, told BIV that he sees himself in that category.
“We don’t have enough data at this point in time,” he said.
“We use customer relationship management (CRM) software.”
Gray told BIV that, in some companies, senior management is to blame for not fully understanding AI technologies and opportunities.
Gray said executives should be putting structures in place to collect data for a future payoff.
AI was top of mind for many retailers at the largest annual retail conference west of Toronto—the November 6 Retail West conference Retail Council of Canada hosted at the Hotel Vancouver.
Speakers on multiple panels discussed ways businesses could use AI, and most attendees raised their hands when asked if they were using AI.
Filter out the people who said that their AI use consisted of making basic searches on platforms such as Google’s Gemini, Perplexity or ChatGPT, however, and it became clear that many retailers have still not started using AI tailored for their businesses.
Sometimes they may be using software that has advanced to include AI features without the business owners realizing it, Gray said.
Other owners of rapidly growing small businesses have made a point to invest in technology that has evolved to incorporate AI.
Vancouver-based pant-seller Duer is one of them.
It opened its eighth store earlier this month and is scouting future store locations in six other cities. About six years ago, it only had one location: its original store on West Hastings Street.
Duer’s wholesales and e-commerce sales back then were similarly a shadow of what they are now.
Customer inquiries in 2018 went to a generic Gmail account that a Duer staffer scrolled through to answer each email manually, Duer’s director of e-commerce and technology Calvin Roex told BIV.
Duer executives then decided they needed to automate.
They started using technology from California-based Gorgias to centralize customer service inquiries and to start to track data, Roex said.
Gorgias helped Duer put a chatbot on its website and have a central point for customer feedback.
Duer workers still answered customer emails manually, but the company started collecting data on which questions customers most asked, and which problems they most experienced.
The technology has since evolved to the point where AI answers Duer’s customers’ emails without first needing human interaction, Roex said.
Similarly, its website’s chatbot is far more sophisticated and can help customers with complex questions about products and problems.
Customers are transferred to human beings on request, if they start to ask the chatbot inappropriate questions or if the chatbot cannot answer the questions, Roex said.
Canadian Tire Corp. Ltd. (TSX:CTC.A) head of AI and emerging technology Cori Covent said on a Retail West panel that her company similarly always has a human overseeing AI activity.
“Anything that might seem that it has a bias built into it, from a question, for example, that might allude to any sort of racism, or anything that might not be consistent with our brand purpose, our brand guardrails will automatically stop that conversation, and it will punt it over to a human to evaluate very quickly,” she said.
Canadian Tire’s chatbot also nudges website visitors with questions, such as where they live, she said.
Knowing that their home is a city that usually only gets slushy snow can determine which kind of tires it recommends.
The chatbot may also ask if the potential customer ever drives to Whistler to go skiing, she said.
The answer to that question may change the chatbot’s recommended product because the driver may be in colder weather or in an environment with more challenging roads.
AI helps retailers manage inventory, stores, marketing
Duer is one of many retailers that are using AI to manage inventory.
Its stores are sometimes used as fulfillment centres, which can complicate inventory flow, Roex said.
“We’re using Toolio software and that is creating some level of AI to predict demand – where inventory needs to be allocated to our retail stores and distribution center,” he added.
Past years’ sales data for a range of products are factored into the mix, as are trends for whether customers are preferring slim-fit or relaxed-fit pants.
Managing inventory well is key for retailers as Vancouver fashion house Lululemon Athletica Inc. (Nasdaq:LULU) found out earlier this year.
Lululemon warned the market in March that it made mistakes with its spring fashion line. The clothing was largely black, white and tan, and lacked colours that consumers wanted to buy. The result was slower sales and higher inventory. It’s share price steadily declined on that news.
The company this summer then saw analysts downgrade its shares because it paused sales of its Breezethrough leggings product line mere weeks after the product’s launch, due to lacklustre sales.
Roex said AI does not yet factor in global fashion data to suggest changes to Duer’s products’ colours or styles, but he said that could be the case in the future.
Determining future store locations has also been done without AI so far. It is largely done by looking at where e-commerce customers are based.
Roex said he can foresee a time when Duer uses AI to take into consideration other factors, such as where complementary retailers are based and where the optimum demographic customers live.
VAMOS Development Advisors principal Michael Penalosa told BIV that Ontario-based PiinPoint has AI-based software that does this already.
“You can’t just say, ‘Look for a store location for a home-appliance company,’ because the AI needs more information,” he said. “It would be garbage in, garbage out.”
When his clients have used PiinPoint he needs data to show where its existing customers live, what they are buying and what their buying patterns and visit frequency is, he said.
He could also provide the company’s sales data, where competition is located and what target customer the company is seeking demographically, he said.
The result is a quick and effective analysis, he added.
Michael Leblanc, who advises retailers and hosts the podcast Voice of Retail, told BIV that many retailers are using AI to create longer descriptions of each individual product on their websites.
“Let’s say you have got 300 products on your website and you want to write descriptions,” he said. “You’re a sole entrepreneur, or you have two people. Start using ChatGPT.”
He said small business owners might get in a rut, and use the same words or phrases were they to write their own copy.
They could instruct ChatGPT to vary the wording and write the marketing material or social media posts, he said.
“Tell me what the best description is of this product, and its benefits, for an e-commerce provider,” he said could be a suggested prompt to an AI platform.
