SingaporeRetail and e-commerce2023 – ongoingZalora
TITAN generative AI platform for search and product tagging
Zalora reported a 4-6% rise in conversion after its OpenAI-based TITAN platform upgraded site search
Problem
Fashion shoppers in Southeast Asia often typed misspelled or loosely worded searches and got no results. Tagging product attributes and writing descriptions by hand also took staff time.
Approach
Zalora, part of Global Fashion Group, built TITAN as an internal AI platform on OpenAI models. The first live uses were search, which handles spelling errors and offers 'Did you mean?' suggestions, and automatic tagging of product attributes such as colour, pattern and sleeve length. Descriptions and SKUs are generated from those tags. A shopping assistant and warehouse forecasting were still in early stages in 2023.
Results
Takeaway
Zalora started with search and catalogue tagging, where results show up in conversion and staff time. Other uses stayed at the experiment stage.
Categories
- Size
- Large enterprise
- Industry
- Retail and e-commerce
- Business function
- Marketing and salesProduction and quality
- Innovation type
- Process innovationProduct and service innovation
- AI technology
- Generative AI
- Adoption
- Partner co-development
- Evidence
- Company's own claim
Sources
Source summaries
Describes TITAN, a platform built with OpenAI. Smart search was live and product attribute tagging was in use. A conversational shopping assistant and warehouse forecasting were still early. It reports a 4-6% conversion rate increase since search launched.
Interview with Zalora's CTO on TITAN. Search accuracy and image-based product tagging were in use, and the tags were more accurate than manual tagging. Conversion rose 4 to 6 percent after launch.
Full details are in the original sources.
