BrazilConsumer goods2024 – ongoingNatura
Busca Inteligente AI product search for beauty consultants
Natura's AI product search served over 1 million consultants and added related products to 99% of searches, per AWS
Problem
Natura's old search matched exact terms only. A consultant who typed a term without an exact match got no results, and a character cap on product names hurt searches for abbreviated product records.
Approach
Natura built Busca Inteligente on AWS in about one month. Amazon OpenSearch Service handles phonetic, vector and text search, and Amazon Bedrock generates tags that link products to related terms. The first release reached 5% of the sales team and was then opened to all consultants in Brazil. Natura plans to extend it to apps and consumer platforms.
Results
Takeaway
Using a language model to tag products fixed the exact-match problem without rebuilding the product catalogue. Natura was still working on measuring whether results lead to carts.
Categories
- Size
- Large enterprise
- Industry
- Consumer goods
- Business function
- Marketing and sales
- Innovation type
- Process innovation
- AI technology
- Generative AIRecommendation and personalization
- Adoption
- Built in-house
- Evidence
- Vendor case study
Sources
Source summaries
The AWS case study describes Natura's search service built on OpenSearch, Bedrock and Lambda. It gives search volume, speed, the share of searches with related products added, and the number of consultants served. The page shows only the year 2024 as its date.
Full details are in the original sources.
