Menu
VACUUM WARS WEBSITE
Vacuum Wars, a leading online platform for vacuum cleaner reviews, faced two key challenges.
Vacuum Wars had amassed a vast and detailed dataset on vacuum cleaners, reviewing each model with meticulous precision. They needed a centralized hub of information to effectively utilize this comprehensive data for marketing and sales through affiliate marketing. This system would not only provide personalized vacuum cleaner recommendations based on user preferences and requirements but also handle user queries related to vacuum maintenance and best practices, ensuring a seamless and informative experience.
The platform required a method to streamline the creation of engaging and consistent scripts for their YouTube review videos. The goal was to ensure that each script matched the unique style and tone of their existing content while also being adaptable for new products.
These challenges called for a sophisticated AI-driven solution to enhance user interactions, sales and optimize the content creation process
To address these challenges, we developed a Chatbot Recommender System and a custom Script-writing Tool.
We designed a chatbot capable of recommending vacuum cleaners from the client’s product database and YouTube review videos. Additionally, it functions as a marketing tool by promoting affiliate product links to drive sales and answering maintenance and best practice queries. Test the chatbot here.
We created a custom script-writing application tailored to Vacuum Wars’ YouTube channel. This application was designed to generate scripts for “Vacuum Review Videos” in the tone and style of the client’s previous content, ensuring consistency and engagement. The application was versatile enough to create content for additional products on which the model was not initially trained on.
The development of the chatbot system involved several crucial steps:
For the script-writing application, we fine-tuned a local LLM using the client’s data. This involved extensive machine learning engineering, including data preprocessing, labeling, and model training to achieve high-quality script generation.
