118 FEATURE data is still largely in the acceleration phase. The technology is ready; the readiness of the organisation is often the bigger hurdle.” Several retailers are using agentic AI tools to support merchandising operations at scale. Walmart has built a suite of AI tools to help buying teams with sales analysis, inventory forecasting and assortment decisions, reducing the time they spend on manual data work. UK retailer Marks & Spencer has worked with Microsoft to deploy a total of 11,000 Microsoft 365 Copilot licences – one for every store manager and support centre colleague. By embedding AI across its systems, Marks and Spencer intends for AI agents to support stock forecasting and ordering, generate marketing materials and power a colleague-help hub. Additionally, the retailer reduced machine learning model execution time by 50 per cent by using Azure, which has directly improved the speed of the M&S Sparks loyalty personalisation engine. US clothing company Guess is piloting Microsoft Copilot Studio-based catalogue enrichment to automate product onboarding, categorisation and metadata normalisation. It aims to transform spreadsheet-based merchandising processes into structured product content capable of supporting discovery, recommendations and a personalised shopping experiences at scale. “What ties these use examples together is the idea of agents taking over the high-volume, data-heavy work that has traditionally taken up so much of a merchant’s time,” says McMahon. “That frees up merchants to focus on the strategic judgement calls and creative decisions that really set a retailer apart. If implemented poorly, AI can dilute merchandising creativity, but the opposite is also true: done well, agentic AI can amplify creativity. The most creative acts in merchandising – such as identifying an emerging trend before the data shows it, curating an assortment that says something meaningful about a brand, building a supplier relationship that unlocks differentiated product – all those things require human judgement, cultural fluency and emotional intelligence, which no AI agent can replicate.” In keeping with its ‘Copilot, not autopilot’ philosophy, Microsoft aims to keep humans at the centre of AI systems. As such, every agentic capability has been designed to generate recommendations and flag low-confidence decisions for humans to review. This philosophy extends to Microsoft’s partner ecosystem, where McMahon believes “the most exciting innovation” is taking place. One partner making notable progress in helping retailers to modernise merchandising is Accenture-Avanade. The company is applying agentic AI to real-time decisions across pricing, inventory, promotions and assortment planning, helping retailers to modernise core merchandising workflows. These Azurebased systems are designed to interpret demand signals and coordinate actions across merchandising functions, enabling merchants to respond faster on product mix, allocation and promotional strategy. Blue Yonder is co-developing the next generation of supply chain planning agents on Azure AI Foundry with Microsoft’s AI Walmart has built a suite of agentic AI tools, including merchant assistant Wally, to reduce the time spent on manual data work
RkJQdWJsaXNoZXIy NzQ1NTk=