Generative AI and Data Analytics-driven Dispute Management for a prominent retailer
Overview.
A large global retailer implemented a system combining Large Language Models (LLMs), Computer Vision, and automated attribute enrichment for item master mapping and matching. This system significantly improved product search, leading to a higher digital order conversion rate and a better customer experience. The system also achieved a massive reduction in the effort required to map and match new products with a success rate exceeding 90%.
Our client, a prominent retailer, has been manually sorting customer complaints received in-store and online, dividing them into one of nine categories. The existing Excel bot built using predefined rules, was only able to accurately categorize these complaints 40% of the time. Long manual processes, limited accuracies, and lack of consolidated view of all complaints led to high inefficiencies in the overall dispute management process.
Solution
Coforge has built an AI/GenAI model to digitize E2E process of categorising the consumer complaints and tag the required vendor supplier
AI/GenAI driven digital process: Consumer raise a complaint -> “Complaint rephrasing -> Case creation -> Complaint categorization -> Supplier identification -> Review and raise dispute”
GenAI / LLM to rephrase the limited words or incomplete complaints to further categorize them under the right dispute category
Identification of Seasonal complaints (e.g. Halloween season) to process them separately
Processing of imbalanced data: due to ambiguities among complaint categories and number of complaints / category
The Impact
100% business rule validation for effective dispute management
Complaint data enrichment for all incomplete complaints
Org level consolidated view of disputes (all stores and online) for further analytics
.jpeg?length=1920&name=CS%20Item%201%20(24).jpeg)
