Overview

This document provides a detailed overview of the AI Insights feature, designed to enhance payroll validations within HRBlizz. The feature utilizes Large Language Model (LLM)-powered enrichment to provide human-friendly summaries and actionable insights relating to Global Payroll Compliance within each regulative region.

It aims to accelerate payroll processing by identifying errors early, reducing manual review time, and improving accuracy. 

Key Themes

• Enhanced Explainability and Actionability: AI enrichment provides clear, business-friendly summaries of anomalies, including “why it’s wrong” and “what to do about it.”
• Efficiency and Reliability: The feature processes concurrently with payroll summary generation, aiming for no additional processing time. It incorporates robust mechanisms like async batching and graceful degradation for reliability.
• Country Agnostic Design: The underlying architecture is designed to be country-agnostic, capable of adapting to various statutory regulations globally where the regulative information is available.
• Phased Rollout and Continuous Improvement: The initial release has certain limitations with plans for future enhancements based on our clients feedback and ever evolving capabilities.

Important Facts

• Core Functionality & Purpose

• • Faster, clearer payroll anomaly detection with human-friendly summaries: The primary goal is to make “payroll data errors” easy to understand and prioritize.
  • Output AI Insights (Results Validation): Activated after payroll processing (at the gross-to-net stage), focusing on “statutory regulation validations”. It feeds recreated payslips from payroll data into an LLM to validate against country-specific statutory regulations.
  • Processing during Payroll Summary Generation: Output AI insights are generated during the payroll summary generation task. “This payroll summary generation is the one task that is generating this AI insights.”  This concurrent task ensures “no additional time will be used to complete the process” to ensure effective quick results. 

• Anonymized Data: When processing data for AI Insights, Mercans employs specific measures to ensure data privacy adhering to our commitment towards ethical and secure AI development :
  • Limited Data Sharing with LLM – Only essential data is shared with the Large Language Model (LLM). This includes the Employee ID (as a string only, with no names), along with anomaly details and aggregate values (such as median or current value).

Exclusion of Sensitive Data Full payroll tables, detailed historical data, and any Personally Identifiable Information (PII) beyond the Employee ID are explicitly not sent to the LLM.

Guardrails for Output – To further protect privacy, a scrub() function is utilized. This function is designed to remove accidental echoes of employee IDs from any text generated by the LLM, acting as a safeguard to prevent PII from being inadvertently exposed in the insights

• Limitations and Future Enhancements
  • Statutory Regulations List: A comprehensive list of output validations is not static as “it’s up to technology to validate and determine an overall regulative assessment”
  • Country Agnostic Approach: The system performs a “Internet search” to first “determine the context of the country and regulation, followed by a thorough validation of these regulations against the clients anonymized data”. This allows it to work across multiple countries (e.g., UAE, Germany, Netherlands). The effectiveness relies on the LLM’s training on specific country knowledge.
  • Performance: While designed for concurrent processing, AI insights generation “takes few minutes” (1.5 to 5 minutes) depending on the complexity. This time is not additive if payroll processing itself is longer.
• Benefits to our Clients:
  • Clarity: Understand detected issues with clear, business-friendly messages and actionable advice.
  • Efficiency: Reduce manual verification time and quicken the identification of critical payroll errors.
  • Accuracy: Leverage advanced AI and rule-based systems to enhance payroll accuracy and compliance.

 

• What does the process look like?

What does it look like in HRBlizz?

Within the normal flow of processing your payroll, when in the result review stage, additional information is displayed as illustrated below, validations are displayed in logical groups and when selected, more detail outlining relevant employees’ insight descriptions and recommended actions will be presented.

 

 

Employees, insight descriptions and recommended actions are visible once expanded.

 

 

Lets have a look in HRBlizz:

 

Important Note

• Even as we incorporate stringent best practices in AI development—including compliance with regulatory requirements for high-risk systems (such as those in the EU AI Act) and fundamental product design principles focused on security, automation, and purpose-built architecture—we firmly assert that human involvement remains a vital component of the payroll processing function.