HR & employee data protection

Last updated: April 2026

"ERPOps monitors your ERP integrations — not your employee data. We watch the pipe. Not the water inside it."

This page explains exactly what data we see, what we never see, and how we protect the boundary for HR teams operating under HIPAA, GDPR, and federal data handling requirements.

This page is provided as a plain-language summary of our practices. Enterprise customers may request our full Data Processing Addendum (DPA), Master Services Agreement (MSA), and security questionnaire responses by contacting sales@erpops.ai.

Real-world HR integration scenarios

Three concrete examples of ERPOps operating on the metadata boundary — diagnosing real problems without ever touching employee data.

PeopleSoft Payroll Integration Failure

Scenario

Your bi-weekly payroll job fails at 11pm. 40,000 employees won't be paid.

What ERPOps sees

PSFT_PAYROLL_BATCH exit code 1, record 18,441 of 40,000, field EMPLID_FK constraint violation.

What ERPOps doesn't see

Employee names, SSNs, bank accounts, or salary amounts.

What happens

AI identifies the foreign key constraint, links it to a known PeopleSoft data integrity issue, surfaces the affected job step, and recommends the SQL fix — without ever seeing who employee 18,441 is.

Workday Benefits Open Enrollment Sync

Scenario

Annual open enrollment. 8,000 employees updating medical, dental, and FSA elections. The Workday-to-carrier sync starts failing midway.

What ERPOps sees

WD_BENEFITS_EXPORT_CARRIER timeout at record 4,203, HTTP 504 from carrier endpoint, retry queue depth 1,847.

What ERPOps doesn't see

Which employees, what plans they selected, PHI of any kind.

What happens

ERPOps detects the carrier endpoint degradation, alerts the team, and triggers a retry playbook — protecting enrollment without exposing a single employee record.

Oracle HCM FMLA Integration

Scenario

Leave management system syncing FMLA designations to Oracle HCM. Intermittent failures are causing leave records to not post, creating compliance risk.

What ERPOps sees

OHC_LEAVE_SYNC_JOB failure rate 12%, average retry count 3.4, error pattern: network timeout on leave_type = [MASKED].

What ERPOps doesn't see

Employee names, medical conditions, physician certifications, or the nature of leave.

What happens

ERPOps identifies the network pattern, correlates it to a known Oracle connector timeout configuration, and surfaces the fix — while the FMLA data itself remains entirely within your environment.