For much of the past two decades, retailers have focused on helping consumers find products online. Search engines became the gateway to discovery, leading to major investments in search optimization, product information, customer reviews and digital advertising.
The underlying assumption was fairly straightforward: if consumers could find your product, you had an opportunity to influence the rest of the decision.
Our research suggests that dynamic is already beginning to change as consumers become more comfortable allowing AI to participate in purchasing decisions. Nearly one-quarter (24%) of Canadians say they have already allowed AI to effectively make a purchase decision for them. A further 37% would be comfortable allowing AI to automatically select products or services based on their preferences.
AI Is Moving From Product Discovery to Purchase Influence
Overall, 54% of Canadians say they trust information from AI platforms, rising to 77% among people who use AI at least weekly.
Consumers are still a long way from handing over every purchase decision to an AI assistant. Shopping can be practical, emotional or simply enjoyable, depending on the person and the category. The larger change is in how much of the work leading up to the purchase may be handed over to AI. Retailers may soon find themselves influencing two decision-makers: the customer and the system helping that customer decide.
The Prompt Could Become Intensely Personal
Today, someone might ask an AI assistant to find the best running shoes for marathon training under $200. I suspect many consumers will eventually maintain a small library of much more detailed shopping prompts across different purchase categories.
A clothing prompt might say:
I like grey, black and neutral colours. Aesthetics are important to me, but I do not like prominent logos. My style is understated rather than trendy. I am willing to pay more for quality, although I tend to regret buying things that require too much maintenance. Recommend clothes that work together and do not duplicate what I already own.
A prompt like that gives AI far more information to work with than a traditional search query, including preferences that may never appear in a product filter. These prompts can continue to improve over time. Consumers could adjust them after disappointing purchases, add photographs for reference or ask AI to remember why one recommendation worked better than another.
Products May be Screened Before the Retailer Gets Involved
The consumer may never browse the category page, notice the promotional banner or compare the products placed beside one another online. The AI assistant could do much of that work first and return with a shortlist.
A brand excluded from that shortlist may have very little opportunity to change the customer’s mind. Retailers will need to understand how AI systems interpret their products. Basic information such as price, availability and technical specifications will still matter. Customer reviews, return policies and previous price changes could also influence a recommendation.
Personal prompts introduce a more difficult layer. Product information does not always explain whether something feels understated, looks more expensive than it is or suits someone who cares about design but avoids anything flashy. Those distinctions are often buried in reviews, images and the language consumers use when talking to one another.
Retailers Need Product Data That AI Can Interpret
Retailers have traditionally developed product information for people and search engines. They may now need to consider whether that information gives AI enough context to match a product with a highly personal request.
This will be especially important when several products satisfy the basic criteria. If five winter coats meet the customer’s budget and technical requirements, what determines which two are recommended? Understanding how those five coats become two recommendations is worth studying now, while AI-assisted shopping is still developing.
Familiarity May Carry More Weight
Personalized AI could make shopping easier, but it may also reinforce existing habits.
If an assistant knows which brands someone buys, which styles they prefer and which previous recommendations worked well, familiar products could become the safer choice. That creates an advantage for established brands. It may also make life harder for challenger brands trying to enter the consideration set.
Consumers will also use AI to discover something new. For challenger brands, getting onto that shortlist may depend on a lower price, stronger reviews or a product feature that fits the customer’s prompt particularly well. .
The implications will vary by category. People may welcome consistency when replenishing household products but expect more variety when buying clothing or planning a trip. Retailers will need to understand when familiarity provides reassurance and when it starts to feel limiting.
Physical Stores Will Have A Different Job To Do
As AI takes more of the work out of routine shopping, physical stores will have to give people a clearer reason to visit.
Their value may increasingly come from helping customers discover preferences they have not yet learned how to describe. A person can touch a fabric, test a product, see how colours work together or ask an employee a question they would not have thought to include in a prompt. They may encounter something that falls outside their established taste and decide they like it anyway.
That experience can then influence future AI-assisted purchases. Someone who discovers a particular fit, texture or design in a store may add it to a saved prompt. The physical store helps the customer learn what they like, and that learning follows them into future shopping decisions.
Stores May Become Places Where Shoppers Refine Their Preferences
More retailers are likely to invest in making stores experiential as a result. That does not require turning every visit into an event. The experience needs to fit the category.
For some retailers, this could mean better access to knowledgeable staff and more opportunities to try products. Others may focus on demonstrations, sampling, customization or spaces that change often enough to reward browsing.
Retailers have spent years making physical stores more efficient purchase channels. Efficiency will remain important, particularly for customers who already know what they want. But stores will carry more responsibility for inspiration, learning and the unexpected parts of shopping.
What Retailers Need To Understand Now
AI shopping will develop at different speeds across categories and customer groups. Consumers may readily delegate the replenishment of household staples while maintaining much more control over clothing, travel or major purchases. Comfort will also change as people gain experience and learn how to give AI better instructions.
Those differences are already something retailers can measure within their own customer base. The immediate research questions include:
- How are customers currently using AI before making a purchase?
- What types of personal preferences are they comfortable sharing?
- Which purchasing decisions are they willing to delegate?
- What determines whether a product makes the AI-generated shortlist?
- Where do customers still value browsing, direct experience or human advice?
- What do people learn in stores that later becomes part of their AI prompts?
By the time these changes become clear in sales data, the behavior behind them may already be well established. Research gives retailers a way to see which customers are changing first, how they’re using AI and where parts of the purchase decision are beginning to happen outside the retailer’s direct influence.
One of the things worth watching is what happens when these two parts of shopping begin feeding each other. A customer might discover a new fit, flavor, product feature or style in person, then add that preference to the instructions they give AI. The next recommendation starts with something the retailer helped the customer discover.
Demetre Eliopoulos brings more than 20 years of experience in market research to his role as Senior Vice President of Research at Angus Reid (angusreid.com). A seasoned expert in polling, Demetre is highly respected for his ability to leverage public opinion to help his clients understand the current trends and forces that shape our landscape. Throughout his career, Demetre has advised hundreds of communications professionals and built insights programs that have yielded impactful results for some of North America’s largest organizations.
