ItalyRetail and e-commerce2025 – ongoingMiroglio Group
Automated product tagging and multilingual descriptions
Miroglio Group cut product-tagging work from 1 month to about 1 hour
Problem
Manual product tagging took time and produced inconsistent categories. Miroglio needed reliable categorization for inventory management and customer browsing.
Approach
Miroglio and Data Reply implemented a tagging system using Claude 3.5 Sonnet on Amazon Bedrock and a machine-learning model trained on historical product images. It analyzed images and product information to generate tags and multilingual descriptions.
Results
Takeaway
Automating catalog work freed staff for creative and strategic tasks. The system combined a general-purpose language model with learning from Miroglio's own images.
Categories
- Industry
- Retail and e-commerce
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Generative AIComputer vision
- Adoption
- Partner co-development
- Evidence
- Vendor case study
Sources
Source summaries
AWS described Miroglio's automated catalog tagging with Data Reply. It reported that tasks previously taking 1 month took approximately 1 hour. The page carries no date; the Internet Archive first saved it on July 10, 2025.
Full details are in the original sources.
