AI Shopping Recommendation Engine for Good Sort

AI-Powered Shopping Experience

Background

WEBSENSA is proud to be a collaborator behind Good Sort, a revolutionary AI-powered shopping platform poised to disrupt the industry.

In today's digital age, information overload is a significant challenge, making it difficult to find reliable product recommendations. Good Sort tackles this issue head-on by leveraging the power of artificial intelligence and expert knowledge.

Understanding the Client

Good Sort aims to revolutionise the shopping experience by providing consumers with personalised, expert-driven product recommendations.

The platform's objective is to address the challenge of information overload, ensuring that users receive reliable and tailored advice for their purchasing decisions. Good Sort envisions a future where shopping is simplified through the power of AI and expert insights.

Project Description‍

WEBSENSA's core contribution to Good Sort lies in developing a robust pipeline for efficient video transcription and annotation. This pipeline plays a critical role in populating the knowledge base that fuels Good Sort's AI experts.

The challenge was to process massive amounts of data from expert podcasts or videos. WEBSENSA's solution involved a two-pronged approach utilising AI models:

1. AI-powered Transcription – Our innovative pipeline utilised an AI model to transcribe the audio content from the videos into text format. This step ensures the rich insights from the experts are captured and preserved for further analysis.
2. Intelligent Annotation – The AI model goes beyond simple transcription by intelligently annotating the data. This crucial step involves identifying speakers within the videos or podcasts, ensuring the RAG system can attribute specific advice to the appropriate expert. This is particularly important for podcasts and videos featuring multiple speakers.

The transcribed and annotated data from the video processing pipeline becomes the cornerstone of Good Sort's AI experts. This rich data allows the AI to understand the nuances of expert recommendations, including specific product features, brand reputation, and comparative analysis between different options.
‍

Good Sort’s Engine Use Case

Imagine you're on Good Sort, seeking the perfect watch. You might ask: "What are some unique Patek Philippe watches suitable for everyday wear?" Good Sort's AI leverages the expert knowledge within the RAG system to deliver personalised recommendations tailored to your query.

This might include insights from renowned watch experts like Yoni Ben-Yehuda or Michael Fossner, alongside a curated selection of Patek Philippe models that align with your search criteria. You even get a direct answer from the "GoodSort collector of expert knowledge" that synthesises the most relevant information from multiple experts.

Managing Big Data

WEBSENSA's contribution goes beyond the video processing pipeline. We also played a vital role in managing the vast amount of product data required for Good Sort's recommendation engine. This involved processing information from thousands of shops, encompassing millions of products. Our expertise in data extraction and transformation ensured this information seamlessly integrated with the RAG system.

Another task connected to data was to identify the occurrence of specific watch models in pop culture and on social media profiles dedicated to this topic. To achieve this, we used available internet sources, where, based on photos and descriptions, we matched watch models to their appearances, e.g., in movies. As part of this process, it was necessary to scrape data from websites and then attempt to assign them to specific watch models.

Results

The collaboration between WEBSENSA and Good Sort yielded impressive results, significantly enhancing the platform's capabilities:

• Efficient Processing of Expert Content: Successfully transcribed and annotated thousands of hours of expert podcasts and videos.
• Enhanced Recommendation Engine: Enabled Good Sort's AI to deliver personalised and expert-driven product recommendations.
• Vast Data Integration: Managed and integrated product data from thousands of shops, encompassing millions of products.

Explore Our Work in Action

See WEBSENSA's AI-driven expert recommendations in action at Good Sort. Visit https://goodsort.ai/ to experience the future of shopping powered by our innovative solutions.

Contact Us

Interested in implementing a similar solution for your business? Contact us – our experts are ready to tailor our offer to your specific needs.