Enhancing Workforce Management: Skello’s AI-Powered Assistant Leveraging Amazon Bedrock
Introduction
Discover how Skello revolutionizes HR software with innovative solutions for employee scheduling and workforce management.
Key Features of Skello
Explore the capabilities that make Skello a leader in HR SaaS, serving diverse sectors while addressing the evolving needs of clients.
Implementing AI for Enhanced User Experience
Learn about the integration of AI to improve data accessibility and user interaction within Skello’s platform.
Navigating Data Protection Challenges
Understand the unique challenges faced in the context of GDPR and multi-tenant environments amidst the deployment of large language models.
Designing the Solution
Get insights into Skello’s architecture and how it leverages advanced language models for seamless querying and visual data representation.
Optimizing Data Schema
Discover the organizational strategies that enhance performance and user experience through a well-planned data structure.
Creating Effective Visuals
Uncover how Skello transforms raw data into meaningful visual presentations automatically, improving comprehension for all users.
Ensuring Security in AI Implementation
Dive into the security measures that protect user data while allowing the benefits of AI-driven insights.
Benefits of the Solution
Witness the impact on user experience and operational efficiency, empowering non-technical users to leverage data effectively.
Conclusion
Reflect on Skello’s journey towards modernizing HR software and enhancing data accessibility, while inviting further exploration of AI applications within the industry.
About the Authors
Meet the minds behind this innovative solution, specializing in AI and software architecture, dedicated to advancing HR technology.
Revolutionizing Employee Management: Skello’s AI-Powered Assistant
This is a guest post co-written with Skello.
In today’s rapidly evolving workplace, effective employee management is more crucial than ever. Enter Skello, a leading HR software as a service (SaaS) solution revolutionizing workforce management across various sectors such as hospitality, retail, healthcare, and construction. With a user base of approximately 20,000 customers and 400,000 daily users throughout Europe by 2024, Skello is continually innovating to meet client needs through features like employee scheduling, time tracking, and payroll preparation.
One of Skello’s latest innovations is the introduction of an AI-powered assistant that enhances user experience and data accessibility. By leveraging Amazon Bedrock, a fully managed service that provides access to high-performing foundation models, Skello ensures robust security and privacy while building generative AI applications.
Navigating the Challenges of Workforce Data
As Skello expanded, it became clear that users needed a more efficient way to access and utilize their workforce data. Many customers, particularly those in HR and operations, found traditional querying tools too technical and time-consuming. This insight led to two key areas of improvement:
-
Quick Access to Non-Structured Data: Users needed to find specific information across various data types—such as attendance logs or performance metrics—without wading through complex queries.
-
Data Visualization: Even with a wealth of data available, users struggled to translate this raw information into actionable insights. Intuitive visual representations are essential for understanding trends without requiring expertise in specialized business intelligence tools.
To address these challenges, we aimed to develop a system that could understand natural language and generate appropriate database queries while ensuring user security and compliance.
Solution Overview: The Power of LLMs
Large Language Models (LLMs) emerged as the ideal solution. Their capabilities in understanding context and natural language make them perfect for transforming user inquiries into precise database queries. Utilizing Amazon Bedrock alongside AWS Lambda, Skello created an architecture that is both scalable and cost-effective, designed to handle the demands of a growing user base.
How It Works
The transformation from natural language to structured queries occurs through a multi-step process:
- User Authentication: Verifying the user’s identity and permissions.
- LLM Processing: Converting the natural language input into a structured query format.
- Query Validation: Ensuring compliance with security policies and user permissions.
- Execution: The database access layer runs the query within the user’s authorized scope.
Complex Queries Made Simple
Beyond basic requests, Skello’s system efficiently handles more complex queries. For instance, when a user asks for "worked hours per week per position for the last 3 months," the system breaks down the request into manageable components:
- Target Metric: Worked hours
- Aggregation Levels: Week, position
- Time Frame: Last 3 months
This structured approach enables the LLM to generate precise calculations without compromising on user intent or data integrity.
Optimizing the Data Schema
A critical component of our system’s effectiveness lies in meticulous data organization. By establishing standardized schema definitions, Skello ensures that similar types of information are consistently stored. For instance, unified date formats prevent confusion, making queries like "Show me events from last week" straightforward.
Additionally, consistently named data fields enhance clarity, allowing the AI to quickly understand user requests like "Show me all full-time employees." This optimization leads to:
- Faster response times
- More accurate answers
- Easier data handling
- Consistent results across different phrasing
Automated Data Visualization
One of the standout features is Skello’s ability to turn raw data into visually compelling charts and graphs automatically. The process includes:
- Smart Label Creation: Automatically labeling axes and titles for clarity.
- Legend Generation: Providing readers with clear explanations of what’s represented in a chart.
- Chart Type Selection: Utilizing the most effective chart types based on the nature of the data.
- Smart Scaling: Ensuring that charts are easy to read and interpret regardless of the data range.
This automated approach demystifies data visualization, making it accessible for all users, regardless of their technical background.
Security-First Architecture
In today’s world, data security is paramount. Skello’s architecture adheres to OWASP best practices, ensuring that security measures are separate from the LLM capabilities. Key components include:
- Role-Based Access Controls: Ensuring users can only access data within their scope.
- Audit Systems: Maintaining logs of all actions to bolster security.
- Multi-Layered Protection: Safeguarding sensitive data through strict validation and partitioning methods.
This architecture not only protects customer data but also enhances user experience by ensuring every interaction is secure and compliant with regulations like GDPR.
Conclusion: Empowering Users Through AI
Through the integration of an AI-powered assistant, Skello is revolutionizing the way HR teams manage their workforce data. By seamlessly transforming queries into structured data and providing intuitive visualizations, Skello empowers users to make informed decisions without needing extensive technical expertise.
As we continue to explore the intersection of AI and human resource management, we invite you to check out the AWS Machine Learning Blog for more insights on AI solutions and their potential applications. If you’re interested in learning more about Skello’s journey towards modernizing HR software, explore our blog series on this transformative journey.
About the Authors
Nicolas de Place is a Data & AI Solutions Architect focused on helping startups harness AI’s potential.
Cédric Peruzzi serves as Software Architect at Skello, where he leads initiatives to enhance software through generative AI features.
If you have questions or suggestions about implementing similar solutions in your multi-tenant environment, we welcome your thoughts in the comments section!