From a meal plan to a supermarket basket with MCP
A concept demonstration of an AI assistant turning meal requirements into recipes, quantities and an itemised basket.
Concept demo: this uses dummy connectors and mock supermarket product data. It is not connected to a live catalogue, basket or checkout.
01Start with the meals you want to make
What if grocery shopping started with the meals you want to make? James Bell uses an MCP-style product connection to illustrate the journey from a customer’s requirements to products in an example catalogue. The demonstration and accompanying guide make the product opportunity concrete.
02Three ideas to explore
- Start with the intended result. The example combines five prepared lunches, dinners for two, dietary preferences and household supplies. The task is to keep those requirements together while turning a plan into a practical shopping list.
- Connect recommendations to available products. A useful retailer integration would need product search, quantities, prices, availability, substitutions and basket management. Recipe generation is one step; proposed items must connect to products the retailer can supply.
- Plan for a live service. A production implementation would need account permissions, customer consent, clear approval, changing-stock handling and a defined checkout route. The guide examines these requirements and the choice between a retailer’s own assistant and a connection to another platform.
This is an illustrative concept, not a completed retail deployment or a measured commercial result.
Read the full conversational-shopping article. For help testing an AI product or connected workflow, explore AI consulting for startups or talk to James.