Gemini Gem: Cart Buddy

Cart Buddy takes grocery optimization to the next level. It's another Gemini Gem I engineered to solve a common problem: where to shop to get the best value for your entire grocery list. Users input their shopping list, and Cart Buddy cross-references prices and availability at a wide variety of local supermarkets. It will ask you for the region and your maximum amount of stores you want to visit. If there are any ambiguous entries on your shopping list, it will also inquire further. Then it provides a strategic recommendation, such as 'Buy these items at Store A and the rest at Store B for a total savings of 15%,' or 'Store C has everything on your list and is the best overall value if you only want to make one stop. You'll always get a clear overview.

Challenges & Solutions

The primary challenge was enabling the AI to handle complex, multi-variable decision-making based on user preferences (e.g., time vs. money). I developed a sophisticated prompt structure that instructs the AI to create a 'decision matrix' internally, weighing factors like total cost, number of store visits, and item availability. This allows it to generate nuanced, personalized recommendations that go far beyond simple price comparison for a single item. Another issue just like with the other Gem is that it limited it's search radius to a very limited amount of stores.