IndiaRetail and e-commerce2024 – ongoingAditya Birla Fashion and Retail
Generative AI product descriptions on a unified data and AI platform
Aditya Birla Fashion and Retail wrote thousands of product descriptions in hours instead of days with LLMs
Problem
Aditya Birla Fashion and Retail (ABFRL) ran about 4,000 retail locations on data that sat in silos and was copied between warehouse, analytics and AI systems. Writing product descriptions for its catalog took days.
Approach
ABFRL moved its data, analytics and machine-learning work onto the Databricks platform. It used large language models to generate product descriptions. It also ran machine-learning models for markdown recommendations, demand forecasting and sentiment analysis of customer feedback.
Results
Takeaway
Bulk text work such as catalog copy was one of the first places a retailer could measure a time saving from language models. The page gave no figures for the quality of the descriptions.
Categories
- Size
- Large enterprise
- Industry
- Retail and e-commerce
- Business function
- Marketing and sales
- Innovation type
- Process innovation
- AI technology
- Generative AI
- Adoption
- Off-the-shelf tool or platform
- Evidence
- Vendor case study
Sources
Source summaries
Databricks describes how ABFRL unified its data, analytics and machine-learning work on one platform. The page lists uses such as markdown recommendations, demand forecasting and LLM-written product descriptions. It reports that thousands of descriptions were created in hours instead of days.
Full details are in the original sources.
