Skello’s AI-Powered HR Assistant: Amazon Bedrock Integration

Skello, a leading HR SaaS solution, enhanced its platform with an AI-powered assistant built using Amazon Bedrock. This addresses the need for improved data access among its 20,000+ customers and 400,000 daily users. The key challenge was providing quick access to diverse data types (employee records, scheduling, attendance) and visualizing this data for analytics, all while maintaining GDPR compliance within a multi-tenant environment. The solution leverages LLMs to translate natural language queries into structured database queries, utilizing AWS Lambda for scalability and cost-effectiveness. A crucial aspect is the robust security architecture, employing OWASP best practices and Amazon Bedrock Guardrails to prevent prompt injection and ensure data privacy. The system features a standardized data schema for efficient search, automatic chart generation with smart labeling and scaling, and granular role-based access controls. Benefits include faster response times, improved accuracy, and simplified data visualization for users with varying technical skills. This implementation showcases how Amazon Bedrock empowers businesses to build secure and user-friendly AI applications while adhering to strict data protection regulations. The architecture includes authentication services, LLM query conversion, query validation (enforcing security policies), and a database access layer operating within user-permitted scopes. Complex queries are handled by extracting query components, generating temporal calculations, and optimizing data schema. The system also employs consistent naming conventions, strategically organized data for fast searches, and specialized data views to pre-aggregate common report requests.

(Source: https://aws.amazon.com/blogs/machine-learning/how-skello-uses-amazon-bedrock-to-query-data-in-a-multi-tenant-environment-while-keeping-logical-boundaries/)