HR & Payroll

An HR and payroll platform covering employee records, payroll runs, leave management, and multi-step approvals, architected around audit logging and compliance requirements rather than adding them after the fact.

Payroll and HR data carry legal and financial consequences that most internal tools aren't built to handle — every change needs a defensible audit trail, and approval workflows need to reflect real organizational policy, not a generic status field. Businesses often outgrow spreadsheet-based HR processes long before they can justify a full enterprise HR suite.

• Build employee management and payroll on an architecture that treats audit logging as a core requirement, not an add-on
• Support configurable, multi-step approval workflows (leave, payroll exceptions) that mirror real company policy
• Give HR teams an AI assistant for routine questions without exposing sensitive employee data beyond its authorized scope
• Keep payroll processing correct and traceable across pay periods

Every write to employee, payroll, or leave records passes through an audit-logging layer that captures who changed what and why. Approval workflows are modeled as configurable multi-step processes rather than a single status field, and an AI HR assistant answers common employee questions against a permission-scoped subset of HR data.

• Built audit logging as a cross-cutting concern applied to every write path, rather than a feature added to individual screens
• Modeled approvals as a configurable workflow engine so different companies' policies (number of approval steps, who can approve what) don't require code changes
• Scoped the AI HR assistant's data access explicitly narrower than a human HR admin's, since its answers are self-service and need a tighter blast radius

• Designing an audit trail detailed enough to satisfy compliance review without slowing down every write operation
• Making the approval workflow engine flexible enough for different company policies without becoming impossible to reason about
• Defining what an AI assistant should and shouldn't be allowed to answer about payroll and personal employee data

• Treated compliance and audit logging as launch-blocking requirements rather than a fast-follow, given the legal exposure of getting payroll wrong
• Kept the AI HR assistant read-only against a narrowly scoped data view, with no ability to modify payroll or employee records

• Designed the audit logging architecture applied across employee, payroll, and leave data
• Built the configurable multi-step approval workflow engine
• Defined the access-scoping model for the AI HR assistant

• Delivered a payroll and HR system with a full audit trail from day one, ahead of compliance review
• Replaced spreadsheet-based leave and approval tracking with a configurable workflow
• Reduced routine HR question volume to human staff via the scoped AI assistant

• Building compliance and audit logging in from the start was significantly cheaper than the alternative of retrofitting it once payroll data already existed in production
• An AI assistant over sensitive HR data needs its access model designed before its conversational behavior — the scoping question comes first