From Guesswork to Personalised Selling
A real-time agentic AI assistant that gives sales staff personalised recommendations and pitch scripts across the entire SKU catalogue.
Customizing AI-supported Inventory Management Solution
Integrating an AI-assisted stock plan recommendation system across a large multi-store retail network.
ThinkCol recently engaged with a multi-store retail chain to harness the power of machine learning to automate stock recommendation across the firm's store network. Before, store and executive managers would rely on business rules in combination with domain knowledge to decide what items to be allocated for each store. Data such as sales forecast, sales performance of similar items across stores, and items affinity were not taken into account.
ThinkCol built an AI matching engine that incorporates big data to match each store with suitable items to maximize profit and diversity. While the engine optimises the stock plan by business-oriented metrics such as ROI, the resulting recommendation plan also has to abide by thousands of business rules that are different for each store.
ThinkCol begins with extensive interviews with users to ensure the understanding of business requirements. Each of these requirements are then mapped into the AI model design to satisfy business rules that cater to specific stores.
Feedback mechanism is also included in the AI model design to address the feedback given by users when they modify the model suggested stock plan and quantities. Inputs to the feedback model include attributes such as user preference, KPI and sales and transaction patterns, and such inputs provide adequate data source for continuous improvements in the AI calculations.
