Case Studies | Retail

GenAI Knowledge Base for Retail Asset Solutions' Audit Staff

About

RAS (Retail Asset Solutions) is a leading provider of inventory management and stocktaking services across the UK and Europe, with audit teams operating out of hours and across many locations.

Challenge

Audit staff needed to reference customer and training documentation mid-audit, but the material was unstructured and slow to search.
That slowed audits down and sometimes pushed staff toward guesswork or leaving items unrecorded.
RAS wanted fast, accurate, compliant access to knowledge in real time, without adding manual upkeep.

Solution

Cloud Combinator built a retrieval-augmented generation workflow on Amazon Bedrock, with an API-driven chatbot (AWS Lambda Function URLs) that retrieves the most relevant material from Bedrock knowledge bases and applies prompt engineering to keep answers accurate and on-brand.

By the numbers:

Real-time, context-aware answers replace manual document searches
Knowledge base stays current automatically as documents change
AWS funding credits offset initial adoption costs

What changed:

Knowledge bases sync automatically whenever documents are uploaded, updated or deleted in Amazon S3, removing manual upkeep entirely.
DynamoDB tracks knowledge-base metadata and sessions for scalability, CloudFormation handles repeatable infrastructure deployment, and CloudWatch gives visibility into workload health.
Security was addressed at every layer, encryption at rest and in transit, least-privilege IAM, and a private-by-default VPC design, and RAS's technical team received runbooks and a knowledge-transfer session to take ownership of the system.