Canadian Tire and Microsoft scale MOSaiC AI platform to sharpen retail timing

Program launch & scope Canadian Tire has moved MOSaiC from test mode into a staged enterprise deployment with Microsoft supplying cloud and model infrastructure. The pilot concluded in 2025 and the company started translating the platform's output into operational plans in 2026. MOSaiC will be rolled into multiple banners and digital touchpoints so decisions can be coordinated across channels. The intention: shift from product-led offers to occasion-led responses.

Data, models and signals The engine fuses first-party loyalty behavior with transaction flows and external context such as weather and local events to form signals about emerging demand. It combines analytics, predictive models and generative features hosted on Microsoft Azure to surface prescriptive recommendations. During testing, the platform flagged over 1,000 distinct life occasions where Canadian Tire’s assets could be relevant. Teams can explore these insights through dashboards and downstream workflows for merchandising and marketing.

Operational effects & enablement Beginning in 2026, merchandising, store and digital teams will use MOSaiC outputs to refine assortments, tailor local experiences and tune personalized promotions. The company is also rolling out AI productivity tools for employees, paired with structured training programs developed with Microsoft and partner business schools. This combination is designed to raise execution speed while preserving governance and security controls. Early platform pilots have informed concrete changes in assortment planning and promotional targeting as the rollout scales.

What this means for customers and staff Customers should see offers and in-store assortments that align more tightly with nearby occasions and short-term demand swings. Employees will access Copilot-style tools to analyze context and make faster, data-backed decisions. The program aims to make local teams more responsive without centralizing every choice, keeping store-level nuance intact.