Alternatives of sold-out items based on AI recommendations
AI implementation and creation of an algorithm to suggest alternative menu items to restaurant customers
Client
Under a non-disclosure agreement (NDA)
Business challenges
- Loss of immediate sales
- Negative impact on customer loyalty
Our solutions
- AI implementation
- Displaying alternatives for sold-out items
- Integration of the solution with the main system
Project description
Initially, we developed a proof of concept (PoC) that leveraged AI to identify the most suitable alternative items on the menu. The engineering team then collaborated with the R&D division to test the developed algorithms and models, ultimately selecting the optimal one.
We implemented a robust backend in PHP to handle data processing and interfacing with the OpenAI and Pinecone APIs. These APIs analyze the attributes of sold-out items and identify suitable alternatives. On the frontend, we used JavaScript, Vue2, and Vuex to retrieve this data via an API and dynamically render alternative products when the original item is unavailable.
To ensure a seamless and responsive user interface, we utilized Sass for styling, enhancing the visual appeal and usability of the product suggestions. This modular approach allowed us to create a highly flexible and scalable solution, easily integrated into the existing platform and adaptable to future requirements.
Technology stack
- JavaScript
- PHP
- Vuex
- Vue2
- OpenAI
- Pinecone
- Sass
Key activities
Results
Platform users gained functionality to protect their sales and increase customer loyalty. Implementing AI has simplified algorithm development and reduced the final cost of implementation
