Supermarket Bot: 3D AI Shopping & Meal Planning Agent

Grocery inflation and household food waste

Food represents one of the largest and fastest-growing expenditures for families in Israel. At the same time, retail pricing across major supermarket chains and localized branches (Rami Levy, Shufersal, Carrefour, and others) is fragmented and volatile. Households face conflicting nutritional requirements - from muscle building and fat loss to low-carb diets, growing children, and general healthy eating - while trying to stick to a realistic budget. Without coordinated planning, consumers overbuy, purchase unaligned ingredients, and discard substantial portions of food each week.

Simultaneous meal planning and cart optimization

Engineering a system that accurately solves grocery budgeting requires addressing complex algorithmic and data hurdles:

• Combinatorial complexity: Optimizing a weekly family meal schedule while minimizing grocery spend is a constrained multi-objective problem. Caloric targets, macronutrient distributions, ingredient overlap, and package sizes must all converge.
• Dynamic branch-level pricing: Supermarket chains update product catalogs and promotional pricing continuously. The solution must standardize heterogeneous catalog feeds and map them to standardized culinary recipes.
• Zero-waste constraints: Buying a carton of eggs or a bag of flour must not leave unused leftovers unaccounted for. Every purchased ingredient must be systematically allocated across the weekly meal schedule.
• Browser-based 3D performance: The user experience needed to be immersive and intuitive without heavy app downloads or hardware barriers, rendering an interactive 3D virtual supermarket store directly inside standard mobile and desktop web browsers.

Supermarket Bot: Agent Shira and 3D Optimization Engine

TendersLab developed Supermarket Bot (סופר-בוט) - an end-to-end intelligent platform that pairs a conversational AI agent ("Shira") with a high-performance optimization solver and an interactive 3D browser environment.

1. Conversational AI agent (Shira)

Shira gathers family profile data: household members, specific nutritional targets (cutting, bulking, low-carb, nutritious maintenance, child growth), and preferred shopping branches. Through friendly conversational interactions, the agent translates complex lifestyle inputs into rigorous mathematical constraints.

2. Joint meal-and-cart mathematical solver

Rather than treating meal selection and grocery shopping as separate tasks, Supermarket Bot's optimization algorithm solves both in a single pass. It selects recipes that satisfy family health targets while choosing exact SKUs and promotions available at the selected local branch to yield the lowest total basket price.

3. Cross-branch regional arbitrage

With a single click, the platform runs regional price comparisons across all competing supermarket branches within a chosen radius. It recomputes the optimal basket and recipe plan for each store, showing consumers exactly where their customized cart is cheapest.

4. Interactive 3D WebGL supermarket

Built with modern WebGL and Three.js, Supermarket Bot provides an engaging virtual grocery store navigation experience right in the browser. Users can explore aisles, visualize their cart dynamically, and inspect products without installing native apps.

Smart nutrition, measurable savings

• Direct consumer savings: Households identify the most cost-effective local branch and promotional bundles for their exact nutritional needs.
• Zero waste design: By coupling recipes with packaged product quantities, every item placed in the basket is consumed throughout the week.
• Seamless accessibility: Completely free and accessible directly via web browser across Israeli food retail networks at supermarket-bot.com.