Case Study - Product Catalog Enrichment with GenAI

The Problem

OnePass aims to provide its members a unified shopping experience across participating brands, including Bunnings, Officeworks, and Priceline Pharmacy, through a single app. Product catalogue data sourced across brands were previously inconsistent and sparse, leading to fragmented product listings, a reduced search experience, and reduced effectiveness of AI-powered product discovery.

The Solution

Vivanti delivered a LLM-assisted product catalogue unification and enrichment solution, modernising a traditionally rule-based problem, by using Google Gemini and BigQuery. LLMs accelerated product attribute extraction, classification, and variant detection, while BigQuery provided the scale to process catalogue data sourced across multiple retailers and brands. Together, LLM-driven enrichment and BigQuery enabled repeatable, production-ready transformations at scale. Alongside reusable ingestion and transformation solutions, Vivanti also established an LLM framework covering prompt management, evaluation, versioning, testing, and output quality.

The Result

The solution delivered a unified and enriched product catalogue with improved standardisation across brand offerings, product taxonomies, and product relationships. Key outcomes included, increased attribute coverage, new structured and searchable attributes enabling richer facets (e.g. dimensions, material, pattern) and enhanced discovery signals such as synonyms, context, use cases, and benefits. Together, these improvements strengthened search relevance and filtering in the OnePass app, and established a foundation for conversational and AI-driven search experiences.