Skip to content
W K
EU AI Act: Not High Risk Q3

Compliance Monitoring Agent

Continuous compliance monitoring - catch gaps before auditors do.

Monitors compliance with labour law, collective agreements, and internal policies. Detects deviations early and escalates before violations.

Analyse your process
Airbus Volkswagen Shell Renault Evonik Vattenfall Philips KPMG

Continuous assessment via AI, classification per rulebook, escalation by severity

The agent assesses operational data from time tracking, payroll and HR via AI detection against rulebooks, classifies deviations deterministically into information, warning or critical violation and routes escalations rule-based to the responsible function.

Outcome: At 234 regulatory changes per day worldwide and new obligations from the EU Pay Transparency Directive, NIS2 and the AI Act, continuous assessment closes the time-lag gap between rulebook change and operational practice.

37% Rules Engine
38% AI Agent
25% Human

The architecture behind this is based on the structural problem of compliance practice:

Eight laws, sixty agreements, new obligations every year

A company with 1,500 employees simultaneously falls under the Working Time Directive, the EU Pay Transparency Directive, GDPR, the EU AI Act, at least one collective agreement, and between 30 and 60 internal policies. Each of these frameworks changes independently. 2025 and 2026 alone bring mandatory electronic time recording in several member states, the Pay Transparency Directive deadline, the AI Act’s phased enforcement, and a wave of national implementation laws. Compliance is not a state. Compliance is a process - and that process has a structural problem.

The Time-Lag Gap

This agent follows the Decision Layer principle: each decision is either rule-based, AI-assisted, or explicitly assigned to a human.

Compliance violations rarely stem from intent. They stem from the gap between the moment a rule changes and the moment operational practice catches up. A collectively agreed pay increase takes effect on 1 April - but the April payroll still runs on the old rates because HR entered the adjustment on 5 April. A new policy on on-call arrangements applies immediately - but the scheduling team learns about it two weeks later.

Thomson Reuters Regulatory Intelligence counted over 61,000 regulatory events globally in 2022 - legislative changes, new directives, regulatory orders. That amounts to 234 regulatory changes per day. In European HR, the volume is lower, but the density is growing. The EU Pay Transparency Directive alone requires a review of all compensation structures, reporting formats, and internal processes by June 2026.

The check on whether all these rules are being followed happens sporadically in most organisations: once a year during the external audit, every few years during a regulatory inspection, on an ad-hoc basis after complaints. Between these checkpoints, months can pass in which deviations exist without anyone noticing.

Rule changes                Day 0
Operational practice drifts  Day 1 - 30
Deviation detected           Day 90 - 365 (next audit)
Correction completed         Day 120 - 400

The time lag is the risk. Not the rule violation itself - that is often trivial to fix. But the fact that it goes undetected for months. A single working time violation can attract fines of up to EUR 30,000 (USD 33,000) in several EU jurisdictions. (UK: penalties vary under the Working Time Regulations 1998, with enforcement by the Health and Safety Executive.) Since the 2022 EU Court of Justice ruling reaffirming time-recording obligations, labour inspectorates across Europe have a sharper instrument at their disposal.

Continuous Monitoring Changes the Structure, Not Just the Speed

The difference between periodic auditing and continuous monitoring is not gradual - it is structural. Periodic auditing examines a sample at a point in time. Continuous monitoring checks every transaction against every relevant rule, every time.

The system consists of four layers:

Rulebook as audit catalogue. Every framework - law, collective agreement, internal policy - is translated into machine-readable checking rules. Not as free text, but as verifiable conditions: “daily working time must not exceed 10 hours” becomes a rule that can be checked against time-recording data. Every rule has a version number and a validity period.

Data integration. Time recording, payroll, HR master data, scheduling - the systems that reflect operational reality are connected as data sources. The agent reads. It does not write.

Deviation classification. Not every deviation carries the same weight. An employee working 15 minutes past the break rule is informational. A systematic working-time violation across an entire department over three weeks is a critical finding. Classification follows a severity matrix:

Severity      Example                                Response
-------------------------------------------------------------------
Information   Single break-time deviation             Log only
Warning       Repeated deviation, one employee        Line manager notified
Critical      Systematic violation, department        HR + Compliance immediately
Escalation    Reportable breach                       Executive + regulator if required

Escalation and follow-up. The escalation matrix determines who is notified at each severity level. Follow-up ensures the corrective action was actually implemented - not just planned. After a defined interval, the system re-checks. Only when the deviation has actually been resolved is the case closed.

Where Accountability Stays

The agent detects deviations. It classifies them. It escalates them. It documents them. It re-checks whether the correction worked. What it does not do: decide what happens next. Whether a working-time violation leads to a formal warning, whether a pay error is corrected retroactively, whether an incident must be reported to a regulator - those are human decisions. And the accountability for the root cause lies with the line manager or the responsible department, not with the individual employee.

This separation is not just a governance choice. It is the reason the system is not classified as high-risk under the EU AI Act. Monitoring and flagging without decisions affecting employment relationships - that is the architecture that enables deployment without a conformity assessment delaying the rollout.

Infrastructure Beyond Compliance

The monitoring engine - versioned rules, operational data checks, deviation classification, escalation, follow-up - is a generic pattern. It does not matter whether the domain is working time, compensation, data protection, or workplace safety. The mechanics are identical.

The re-check pattern (was the deviation actually resolved?) is needed by every agent that initiates corrective processes. The Audit Agent needs it for open remediation items. The Onboarding Agent needs it for mandatory training. The Payroll Agent needs it for retroactive calculations.

And the Audit Trail that monitoring generates as a by-product - when a deviation was detected, who was notified, what action was taken, when the re-check happened - is exactly the documentation that external auditors and regulatory inspectors expect as evidence. Audit preparation shrinks from weeks to hours because the evidence already exists.

When the Switch Pays Off

The direct calculation is straightforward. An HR team that currently spends 15 percent of its time on manual rule monitoring and retrospective deviation correction reclaims that time - not entirely, because the corrective actions themselves remain human work, but the detection and documentation effort largely disappears.

The real calculation is different. A single working-time violation can attract fines of up to EUR 30,000 (USD 33,000) - per violation, not per incident. If a department of 40 employees systematically exceeds working-time limits over three months and the labour inspectorate discovers it during an inspection, the exposure is not a single fine. Add legal costs, back payments, and the reputational damage with the works council and workforce.

Continuous monitoring replaces that risk calculus with a defined process: detect, classify, escalate, correct, re-check. Every day, not once a year.

Micro-Decision Table

Who decides in this agent?

8 decision steps, split by decider

37%(3/8)
Rules Engine
deterministic
38%(3/8)
AI Agent
model-based with confidence
25%(2/8)
Human
explicitly assigned
Human
Rules Engine
AI Agent
Each row is a decision. Expand to see the decision record and whether it can be challenged.
Define compliance indicators Establish measurable compliance checks per policy and regulation Human

Compliance indicators defined by legal, HR, and compliance teams

Decision Record

Decider ID and role
Decision rationale
Timestamp and context

Challengeable: Yes - via manager, works council, or formal objection process.

Collect monitoring data Pull relevant data from HR systems for compliance checking AI Agent

Automated data collection from defined sources

Decision Record

Model version and confidence score
Input data and classification result
Decision rationale (explainability)
Audit trail with full traceability

Challengeable: Yes - fully documented, reviewable by humans, objection via formal process.

Evaluate compliance status Check data against defined rules and acceptable ranges Rules Engine

Deterministic rule application per compliance indicator

Decision Record

Rule ID and version number
Input data that triggered the rule
Calculation result and applied formula

Challengeable: Yes - rule application verifiable. Objection possible for incorrect data or wrong rule version.

Detect deviations Identify out-of-range values or policy violations Rules Engine

Threshold comparison and rule violation detection

Decision Record

Rule ID and version number
Input data that triggered the rule
Calculation result and applied formula

Challengeable: Yes - rule application verifiable. Objection possible for incorrect data or wrong rule version.

Alert responsible parties Notify compliance officer and process owner of detected deviation Rules Engine

Escalation rules based on deviation type and severity

Decision Record

Rule ID and version number
Input data that triggered the rule
Calculation result and applied formula

Challengeable: Yes - rule application verifiable. Objection possible for incorrect data or wrong rule version.

Track remediation Monitor corrective actions to completion AI Agent

Automated tracking with deadline monitoring

Decision Record

Model version and confidence score
Input data and classification result
Decision rationale (explainability)
Audit trail with full traceability

Challengeable: Yes - fully documented, reviewable by humans, objection via formal process.

Review remediation effectiveness Verify that corrective action resolved the compliance gap Human

Human verification that root cause has been addressed

Decision Record

Decider ID and role
Decision rationale
Timestamp and context

Challengeable: Yes - via manager, works council, or formal objection process.

Generate compliance reports Produce compliance status reports for stakeholders AI Agent

Automated reporting per stakeholder and regulatory requirements

Decision Record

Model version and confidence score
Input data and classification result
Decision rationale (explainability)
Audit trail with full traceability

Challengeable: Yes - fully documented, reviewable by humans, objection via formal process.

Decision Record and Right to Challenge

Every decision this agent makes or prepares is documented in a complete decision record. Affected employees can review, understand, and challenge every individual decision.

Which rule in which version was applied?
What data was the decision based on?
Who (human, rules engine, or AI) decided - and why?
How can the affected person file an objection?
How the Decision Layer enforces this architecturally →

Does this agent fit your process?

We analyse your specific HR process and show how this agent fits into your system landscape. 30 minutes, no preparation needed.

Analyse your process

Governance Notes

EU AI Act: Not High Risk
Not classified as high-risk under the EU AI Act - the agent monitors process compliance, not employee behaviour. The distinction between process monitoring and employee surveillance must be clearly defined and maintained. Works council co-determination rights apply to the introduction of monitoring systems. The scope of what is monitored, how deviations are handled, and who receives alerts must be documented in works council agreement. GDPR applies to any personal data processed in the monitoring.

Assessment

Agent Readiness 64-71%
Governance Complexity 51-58%
Economic Impact 58-65%
Lighthouse Effect 41-48%
Implementation Complexity 44-51%
Transaction Volume Daily

Prerequisites

  • Defined compliance indicators per policy and regulation
  • Data access to HR systems being monitored
  • Compliance officer assignment per domain
  • Remediation tracking infrastructure
  • Reporting templates for regulatory and audit purposes
  • Works council agreement on automated compliance monitoring scope

Infrastructure Contribution

The Compliance Monitoring Agent builds the continuous monitoring infrastructure that supports all governance-intensive agents. The deviation detection, remediation tracking, and compliance reporting patterns established here are the operational governance layer that high-risk agents (Candidate Screening, Performance Review, People Analytics) depend on. Builds Decision Logging and Audit Trail used by the Decision Layer for traceability and challengeability of every decision.

What this assessment contains: 9 slides for your leadership team

Personalised with your numbers. Generated in 2 minutes directly in your browser. No upload, no login.

  1. 1

    Title slide - Process name, decision points, automation potential

  2. 2

    Executive summary - FTE freed, cost per transaction before/after, break-even date, cost of waiting

  3. 3

    Current state - Transaction volume, error costs, growth scenario with FTE comparison

  4. 4

    Solution architecture - Human - rules engine - AI agent with specific decision points

  5. 5

    Governance - EU AI Act, works council, audit trail - with traffic light status

  6. 6

    Risk analysis - 5 risks with likelihood, impact and mitigation

  7. 7

    Roadmap - 3-phase plan with concrete calendar dates and Go/No-Go

  8. 8

    Business case - 3-scenario comparison (do nothing/hire/automate) plus 3×3 sensitivity matrix

  9. 9

    Discussion proposal - Concrete next steps with timeline and responsibilities

Includes: 3-scenario comparison

Do nothing vs. new hire vs. automation - with your salary level, your error rate and your growth plan. The one slide your CFO wants to see first.

Show calculation methodology

Hourly rate: Annual salary (your input) × 1.3 employer burden ÷ 1,720 annual work hours

Savings: Transactions × 12 × automation rate × minutes/transaction × hourly rate × economic factor

Quality ROI: Error reduction × transactions × 12 × EUR 260/error (APQC Open Standards Benchmarking)

FTE: Saved hours ÷ 1,720 annual work hours

Break-Even: Benchmark investment ÷ monthly combined savings (efficiency + quality)

New hire: Annual salary × 1.3 + EUR 12,000 recruiting per FTE

All data stays in your browser. Nothing is transmitted to any server.

Compliance Monitoring Agent

Initial assessment for your leadership team

A thorough initial assessment in 2 minutes - with your numbers, your risk profile and industry benchmarks. No vendor logo, no sales pitch.

30K120K
1%15%

All data stays in your browser. Nothing is transmitted.

Frequently Asked Questions

Is this agent employee surveillance?

No. The agent monitors process compliance - whether organisational processes operate within defined parameters (working time limits, approval workflows, data retention rules). It does not track individual employee behaviour. The distinction is fundamental and explicitly defined in the works council agreement.

Who decides what compliance indicators are monitored?

Compliance indicators are defined collaboratively by legal, HR, and compliance teams, with works council consultation where co-determination applies. The agent executes monitoring against defined rules - it does not decide what to monitor.

What Happens Next?

1

30 minutes

Initial call

We analyse your process and identify the optimal starting point.

2

1 week

Discover

Mapping your decision logic. Rule sets documented, Decision Layer designed.

3

3-4 weeks

Build

Production agent in your infrastructure. Governance, audit trail, cert-ready from day 1.

4

12-18 months

Self-sufficient

Full access to source code, prompts and rule versions. No vendor lock-in.

Implement This Agent?

We assess your process landscape and show how this agent fits into your infrastructure.