51³Ô¹ÏÍø

AI in retail: Insights from the 2025 Reshaping the Future of Retail Conference

The 2025 Reshaping the Future of Retail Conference convened over 250 Canadian retail leaders, academics, and students at the Sofitel Montreal Golden Mile hotel on Friday, September 26. Hosted by the Bensadoun School of Retail Management and powered by KPMG, the program featured a keynote by L’Oréal Canada CEO An Verhulst-Santos and a strong roster that included Anindya Ghose professor at New York University and author of Tap and Thrive; Hélène Drolet, vice-president of operations North America at Circle K/Couche-Tard; and James Clark, 51³Ô¹ÏÍøengineering professor and co-lead of the Retail Innovation Lab.

An Verhulst-Santos
An Verhulst-Santos (President & CEO, L’Oréal Canada)
The day prioritised practicality and community. Speakers showed where artificial intelligence (AI) already fits inside real-world operations and offered actionable guardrails for responsible implementation. Assumptions were tested, best practices exchanged, and the Canadian retail community arguably left more connected and better equipped for what’s next.

Across sessions, several ideas kept resurfacing. Three in particular framed how retailers might create value today while building capabilities for what comes next.

Personalization for decision support

Personalization surfaced repeatedly from the lens of helping customers make decisions with less effort. In the beauty sector, L’Oréal’s technology suite, including La Roche-Posay’s SPOTSCAN+ and Cell BioPrint skin analysis, narrows a sprawling aisle into a regimen tailored to an individual’s needs.

On a separate track, Alain Tadros, chief marketing officer and head of digital strategy at Metro, described a loyalty strategy built on personalization. AI is expected to scale relevance, cadence, and offer quality for Moi members. In Metro’s case, personalized experiences reportedly delivered 10 times stronger sales performance than non-personalized approaches, reinforcing the bet.

Eric Morris, managing director of Google Canada, added useful context: Canadian shoppers may be better informed yet more fatigued, which suggests clarity, not spectacle, is what resonates. Taken together, these perspectives imply that AI can translate data into timely, human-sensible guidance across channels.

Key takeaway: AI functions as a force multiplier for relevance, pushing personalization to a level that feels valuable, earned, and welcome.

Speakers on stage with mics
Vilma Todri, James Clark, Alain Tadros

Always start with good data and governance

Many retailers appear to be awash in data yet short on decision-grade inputs. Speakers argued that the most important work happens upstream: repairing historical setups, enforcing standard operating procedures (SOPs) and naming conventions, defining ownership, and keeping datasets current.

Prof. Ghose reinforced the point: despite abundance, only a small slice of enterprise data typically informs decisions. He advised to allocate the majority of effort to data cleanliness and readiness, supported by organizational scaffolding: dedicated teams to set policy, integrate tools into workflows, and align AI initiatives to the right business problems.

Governance, in that sense, isn’t window dressing; it’s the guardrail that ties the tech stack to outcomes, limits tool sprawl, and makes ethics and measurement explicit. Models improve quickly; messy pipelines usually don’t.

Key takeaway: Treat data stewardship and governance as operating disciplines, not short-term projects. Without clean inputs and clear guardrails, AI pilots stall long before they create value.

Vilma Todri, Nicolas Hien, Anindya Ghose
Vilma Todri, Nicolas Hien, Anindya Ghose

Improving daily efficiency

A recurring theme was AI’s role in streamlining everyday work. At Circle K/Couche-Tard, labour planning shifted from a time-consuming manual task to an automated workflow that ingests store attributes, transaction history, local events, weather, and multi-year sales. The result was tighter schedules that freed managers from administrative drudgery and returned more time to the floor.

Dollarama offered a related people-operations example: implementing AppyHere to modernize frontline hiring reportedly reduced time-to-hire from 17 days to about 1.17 days. The ApplyHere AI allowed Dollarama to reallocate $10 million in hiring administration costs to customer-facing work and decreased employee turnover by 35 per cent.

On the marketing side, creative operations are also evolving. Research shared by Prof. Vilma Todri (Emory University) indicates scenarios where GenAI-generated ads outperformed expert-made creatives on purchase intent—useful in categories where media assets decay quickly. This doesn’t sideline human judgment; it gives teams a way to keep up with content velocity while acting as editors-in-chief to ensure brand standards are protected.

Key takeaway: AI’s most reliable value showed up as reclaimed time and better resource allocation: automate repeatable tasks, redeploy hours to higher-impact work, and maintain human oversight where judgment matters.

Looking forward

Retail now operates in a VUCA world, characterized by volatility, uncertainty, complexity, and ambiguity, where leaders, professionals, faculty, and students face pressure to make “right†decisions faster while under greater scrutiny and with fewer certainties.

Events like this help by building a community of practice - a place to share ideas, compare evidence, and surface risks before projects scale. Threaded through the day was a simple reminder: AI sits downstream of humans. We choose the data, set the objectives, design the guardrails, and live with the results. Used with care and precision, AI can elevate service and decision quality; used poorly, it can just as quickly scale error.

On a personal note, as a young professional at the start of my career, I left inspired and reassured by how deliberately retail leaders are approaching AI. In a news cycle that can make the future feel intimidating, it was heartening to see rigour, care, and human judgment prioritized. AI did not read as an adversary here; it felt like a partner that, when guided well, can help people do their best work.

Acknowledgement: I’d like to thank my friend and classmate, Sini Yang, whose diligent note-taking during the conference helped support the creation of this article. Thank you, Sini!


Bensadoun School of Retail Management

To stay engaged with this work, and help shape what comes next, explore upcoming events, research, and industry collaborations at the Bensadoun School of Retail Management.

Learn more about the Master of Management in Retailing (MMR).

Feedback

For more information or if you would like to report an error, please web.desautels [at] mcgill.ca (subject: Website%20News%20Comments) (contact us).

Join us for our next session. 

Our information sessions provide an excellent opportunity to learn more about the program’s curriculum and requirements, while also offering a chance to connect with program specialists. 

Back to top