AI in Retail

Retailers and e-commerce platforms applying AI to personalisation, merchandising, pricing and store operations.

Use cases

24/7 conversational shopping and customer service agents Emerging

Problem:: Retailers need round-the-clock, multilingual customer support and shopping assistance without proportionally scaling human staff.
Capability:: Conversational agents / real-time voice AI
Value:: Deployments like Yamada Denki's GPT-Realtime agent reached 30,000 shoppers in two weeks with 92% positive experience, showing potential for scaled, low-cost customer engagement.

Barriers: Trust and escalation paths to humans; documented consumer frustration when bots block access to human agents in complex cases

AI-driven demand forecasting for inventory and pricing Emerging

Problem:: Retailers struggle to accurately predict store-level and SKU-level demand across seasons, promotions, and store variations, leading to stockouts or excess inventory.
Capability:: Time-series forecasting with foundation models (e.g., TimesFM) including anomaly detection and covariate handling
Value:: Improved forecast accuracy over seasonal naive and classical baselines can reduce inventory holding costs and stockouts across multi-store operations.

Barriers: Data quality and integration across store systems; need for retraining/monitoring against real-world drift

AI-powered product discovery and search Experimental

Problem:: Traditional e-commerce search relies on seller-provided labels and keyword matching, often failing to surface the most relevant products and hurting conversion rates.
Capability:: Neurosymbolic search / knowledge-graph reasoning over product attributes
Value:: More accurate, inspectable product discovery can significantly lift search relevance and conversion versus incumbent search engines, as shown by benchmark wins against major platforms.

Barriers: Integration with existing catalog and search infrastructure; need for structured, high-quality product attribute data

Virtual try-on and 3D product visualization Emerging

Problem:: High return rates in fashion e-commerce stem from customers being unable to assess fit, size, or appearance before purchase.
Capability:: Generative AI / computer vision for 3D model generation from product photos
Value:: Retailers like ASOS, Breuninger, and Maybelline report measurable reductions in return rates after moving virtual try-on from pilot to production.

Barriers: Image data quality and pipeline costs; integration with existing product catalogs