Close
Beauty Square Deploys Generative AI Chatbot on AWS to Transform Internal Knowledge Access
Beauty Square is a fast-growing cosmetics and skincare company headquartered in East Africa, serving customers through both retail outlets and online platforms. With increasing workforce size and a broad range of product offerings, Beauty Square operates in a dynamic and information-intensive environment, where accurate and timely internal communication is essential for scaling operations, maintaining compliance, and delivering exceptional customer experience.
As Beauty Square scaled operations across multiple departments and geographies, employees encountered growing difficulties in accessing institutional knowledge—especially HR policies, operational guidelines, and product documentation.
Decentralized Documentation
Important documents such as leave policies, SOPs, and compliance protocols were spread across shared drives, emails, and legacy portals. Employees wasted time locating the most current information, leading to errors and duplication of effort.
Inefficient Internal Query Handling
HR and IT teams were overwhelmed by recurring queries from employees. Without an automated system, valuable time was spent responding to questions that could be self-served with the right tools.
Onboarding Bottlenecks
New employees struggled to find information, delaying onboarding and impacting time-to-productivity.
Lack of Intelligent Search
Keyword-based search tools proved inadequate in understanding natural language queries, especially when phrased in conversational formats.
Risks if Unaddressed:
Reduced employee productivity
Poor internal satisfaction and engagement
Delayed compliance responses
Increased operational overhead for HR and admin teams
To address these challenges, Kitsilano Technologies Limited, an AWS Advanced Consulting Partner with specialization in AI/ML and cloud-native workloads, designed and implemented a Generative AI-powered Enterprise Chatbot solution leveraging a Retrieval-Augmented Generation (RAG) architecture on AWS.
The solution combined document ingestion, embedding generation, semantic vector search, and large language model-based response generation in a secure, scalable, and serverless architecture:
A custom chat web app was developed for employee use via desktop and mobile.
Integrated with Amazon Cognito for secure SSO and multi-factor authentication.
JWT tokens ensured every user request was validated for enterprise access.
Enterprise documents (PDFs, Word files) were uploaded to Amazon S3.
Amazon SageMaker Processing Jobs:
Parsed and cleaned the content.
Generated text embeddings using pre-trained foundation models.
Stored embeddings in Amazon OpenSearch Service (Vector DB) for high-speed vector similarity search.
On user query, the following sequence occurred:
a. Request Handling:
The frontend sent the query to Amazon API Gateway, passing the JWT token.
API Gateway validated the token and forwarded the request to AWS Lambda.
b. Document Retrieval:
Lambda queried OpenSearch for top-k similar documents based on the query’s vector representation.
c. LLM Response Generation:
The matched documents and user query were passed to a Foundation Model via Amazon Bedrock.
Bedrock generated a natural-language answer contextualized to Beauty Square’s internal documentation.
d. Response Delivery:
The response was returned via Lambda and API Gateway to the frontend chat interface.
IAM policies controlled access to services and resources.
Amazon CloudWatch for performance metrics and logs.
Amazon KMS ensured encryption at rest and in transit.
CloudTrail provided audit logs for all API access and administrative actions.
The chatbot was deployed across the organization and made available to all authenticated staff via web and mobile apps.
✅ 60% reduction in internal HR and IT ticket volume
📉 40% faster onboarding for new employees through self-service policy access
💬 85%+ of employee queries answered directly by the chatbot without human intervention
🔄 Zero downtime, with the architecture scaling seamlessly across usage spikes
🔐 100% compliance with internal security standards and enterprise authentication policies
Boosted employee satisfaction and engagement
Reduced operational burden on support teams
Enabled a foundation for future AI capabilities like voice queries and multilingual support
Positioned Beauty Square as a digitally mature workplace embracing intelligent automation
Kitsilano Technologies Limited is an Advanced AWS Partner based in Kenya, with validated Service Delivery credentials for Amazon RDS, Amazon Glue, QuickSight, and Systems Manager. Kitsilano is the only AWS Partner in East Africa with the SAP Competency, and a recognized leader in helping businesses transform operations using Generative AI, Machine Learning, and Cloud-Native architectures.
The team brings deep expertise in data pipelines, LLM applications, intelligent search, and enterprise-grade deployments on AWS.
Customer: Beauty Square (Retail and Cosmetics)
Challenge: Disconnected document systems, employee productivity losses, HR workload
Solution: GenAI-powered Enterprise Chatbot with RAG architecture on AWS
AWS Services Used: Amazon Bedrock, SageMaker, OpenSearch, Lambda, S3, Cognito, API Gateway
Outcomes: +85% auto-resolution, 60% fewer tickets, improved compliance & onboarding
Partner: Kitsilano Technologies Limited (Advanced AWS Partner, East Africa)
https://drive.google.com/file/d/1tFCMF46ZdUwrDdSsNtjyCWnCk1NmyT-j/view?usp=sharing
